Knowledge catalog
Generated referenceRead-only pages rendered from the Quake AI knowledge graph.
Concepts
185 concept knowledge nodes across Quake AI documentation.
See also the editorial glossary for authored prose definitions.
- →
Access Control
Authorization rules decide which identities may perform which actions on which resources after authentication succeeds. Role assignments, policy files, and API scopes combine to enforce least privilege across projects and services. Auditors review access matrices when onboarding teams or launching regulated workloads.
9 doc pages
- →
Account
A customer record ties billing profiles, contracts, support entitlements, and default projects under a provider. Usage rolls up to accounts for invoicing while day-to-day API work stays project-scoped. Operators use accounts to separate subsidiaries or resellers without merging their networks.
2 doc pages
- →
Affinity
A Nova server-group policy schedules instances onto the same physical host when capacity allows, improving locality. Anti-affinity spreads members across hosts to reduce correlated hardware failure. You pick the policy when creating the group and launching members.
3 doc pages
- →
AI Agent
A runtime combines models with tools, memory, and guardrails to pursue multi-step goals across conversation turns. Agents plan actions, call HTTP or CLI tools, then summarize outcomes instead of emitting single replies. Hosting on VMs keeps connectors near private APIs and regulated data.
3 doc pages
- →
Ai Tools
Structured function interfaces let models invoke HTTP calls, SQL, or shell commands with validated JSON arguments. Tool schemas separate read-only inspectors from mutating operations and cap side effects. Operators log tool calls for security review the same way they audit API keys.
9 doc pages
- →
Ansible
13 doc pages
- →
Anti Affinity
A Nova server-group policy places each member on a different hypervisor when the scheduler can satisfy the rule. The pattern limits blast radius when a host fails or maintenance drains a rack. Soft variants prefer separation without hard failures when capacity is tight.
1 doc page
- →
API
A programmatic contract exposes create-read-update-delete flows over HTTP or RPC with stable error models. Clients automate what the console shows, using pagination, idempotency keys, and rate limits for bulk work. Version headers or paths pin behavior across upgrades.
29 doc pages
- →
Api Tokens
1 doc page
- →
Api Versioning
Providers evolve payloads and routes without instantly breaking every client by advertising supported ranges. URI paths, Accept headers, or microversions let callers opt into new fields while legacy traffic keeps working. Sunset dates and changelogs give integrators time to migrate automation.
1 doc page
- →
Application Credentials
4 doc pages
- →
Apps
1 doc page
- →
Authentication
The process proves who you are using passwords, API keys, WebAuthn, or federated SAML and OIDC exchanges. Keystone validates credentials and issues tokens other services trust. Pair strong authentication with short-lived tokens and rotation for service accounts.
8 doc pages
- →
Automation
Pipelines, scripts, and infrastructure code provision, configure, and remediate cloud resources on every change. Git-reviewed automation replaces repetitive console work and captures intent for auditors. Good automation wires secrets safely, supports dry runs, and surfaces actionable failures.
52 doc pages
- →
Availability
1 doc page
- →
Backup
A point-in-time copy of data or configuration enables restore after corruption, operator error, or ransomware. Policies define frequency, retention, encryption, and off-site replication targets. Untested backups are guesses; restore drills validate RTO and RPO claims.
9 doc pages
- →
Billing
Metering turns resource usage such as vCPU hours, gigabytes, and egress into invoice line items. Project tags feed chargeback while accounts roll up totals for finance systems. Transparent calculators help engineers estimate costs before shipping features.
11 doc pages
- →
Block Storage
Fixed-size blocks with stable addresses back virtual disks, databases, and VM root volumes in SAN-style layouts. Cinder drivers attach volumes over iSCSI or similar transports with snapshot and clone workflows. Performance tiers map to underlying media and replication policies.
32 doc pages
- →
Buckets
Named containers in object storage hold arbitrary blobs keyed by path-like names plus optional version IDs. Policies control encryption, ACLs, lifecycle transitions, and public access. Treat buckets as flat namespaces rather than POSIX directories to avoid hot-prefix bottlenecks.
5 doc pages
- →
Certificates
X.509 documents bind DNS names or service identities to public keys for TLS handshakes and optional mutual TLS. Operators renew before expiry to avoid browser warnings and broken APIs. Private keys stay on instances or self-managed reverse proxies while public chains validate trust.
2 doc pages
- →
Changelog
A chronological record highlights features, fixes, deprecations, and breaking API shifts between releases. Integrators read it before upgrading SDKs or Terraform providers. Linking issues and migration guides turns the changelog into an operational checklist.
2 doc pages
- →
Cicd
5 doc pages
- →
Clones
5 doc pages
- →
Cloud Init
First-boot agent that applies user-data scripts when an instance launches. See the cloud-init concept page for formats, debugging, and launch methods.
13 doc pages
- →
Cloud Native
Design favors elastic scaling, health endpoints, immutable deploys, and twelve-factor style configuration via env vars. Containers and Kubernetes appear often, yet VMs remain vital for stateful databases and compliance boundaries. Observability is built-in rather than bolted on after incidents.
1 doc page
- →
Cloud Native Computing
1 doc page
- →
Cloud Terminology
Shared vocabulary maps product names to OpenStack services and industry standards so docs stay precise. Consistent terms reduce support churn when engineers compare runbooks across teams. Glossaries anchor autolinks from prose to canonical definitions.
1 doc page
- →
Clouds Yaml
Configuration file lists named clouds with auth plugins, regions, interface URLs, and project defaults. OpenStackSDK and OSC honor the file so operators switch contexts with environment variables or flags. Keep secrets out of git and restrict filesystem permissions on the file.
2 doc pages
- →
Cluster Templates
Magnum cluster templates describe Kubernetes versions, network drivers, flavors, labels, and add-on charts for worker pools. Templates are part of the Magnum API; OpenTofu modules are an alternative for defining cluster topology as code. Standard templates encode security baselines every team inherits.
9 doc pages
- →
Clusters
Coordinated node groups such as Kubernetes control planes and workers or replicated databases operate as one unit. Autoscalers grow or shrink membership based on signals you define. Network policies and security groups fence east-west traffic inside the cluster edge.
15 doc pages
- →
Compute
Hypervisor-backed services schedule virtual machines, enforce quotas, and expose power, console, and migration APIs. Operators patch kernels, evacuate failed nodes, and tune CPU oversubscription carefully. Features like NUMA hints and huge pages target latency-sensitive workloads.
34 doc pages
- →
Configuration Management
13 doc pages
- →
Console
The web UI browses resources, shows metrics, launches instances, and surfaces organization settings. Role assignments gate which projects appear in the session. Consoles complement CLIs for exploration while automation still flows through APIs.
39 doc pages
- →
Console Apps
4 doc pages
- →
Containers
OCI-packaged processes share the host kernel yet isolate filesystems, cgroups, and namespaces. Orchestrators schedule replicas, mount secrets, and roll out image updates safely. Containers boot faster than VMs but trade some isolation for density.
70 doc pages
- →
Cors
1 doc page
- →
Credentials
Secrets such as passwords, API keys, kubeconfigs, and cloud signing pairs authenticate clients to services. Never log or commit them; rotate on compromise and scope to least privilege. Identity systems issue short-lived variants where possible to shrink blast radius.
17 doc pages
- →
Dashboard
1 doc page
- →
Databases
Managed or self-run engines store relational, document, wide-column, or time-series data accessed over connection strings. Persistent disks or object stores back durability while replicas handle failover. Capacity planning pairs vCPU, RAM, and IOPS with backup and patching runbooks.
15 doc pages
- →
Debugging
Engineers reproduce defects, inspect logs and metrics, trace requests, and validate fixes under load. Correlation IDs stitch user journeys across microservices. Structured hypotheses beat random toggles when outages pressure timelines.
2 doc pages
- →
Definitions
Short glossary entries explain terms consistently across docs, UI strings, and support macros. Autolinking surfaces definitions without duplicating paragraphs in every article. Updating entries after rebrands keeps assistants and humans aligned.
1 doc page
- →
Deployments
1 doc page
- →
Disks
Reader language for attached block storage on a VM often says disk or EBS volume rather than Cinder volume. Quake AI persists data on block storage volumes you attach to instances with optional snapshots and clones.
1 doc page
- →
Dnat
Destination NAT rewrites inbound packet addresses so public listeners reach private instance ports. Floating IPs rely on DNAT at the edge while connection tracking maps return flows. Misrules cause one-way TCP hangs or blackholed UDP.
1 doc page
- →
DNS
The distributed naming system maps hostnames to A, AAAA, SRV, CNAME, and TXT records resolvers cache hierarchically. Split-horizon views serve different answers inside VPCs versus the public internet. Low TTLs speed cutovers; stale caches delay migrations after IP changes.
9 doc pages
- →
Docker
The container engine builds images from Dockerfiles, publishes to registries, and runs processes with cgroup limits. Compose files model multi-service dev stacks that mirror production dependencies locally. Harden daemons, pin base images, and scan layers just like VM images.
39 doc pages
- →
Docker Compose
YAML declares services, networks, volumes, and env files so one command starts interdependent containers. Compose accelerates integration tests before Kubernetes manifests ship. Manage secrets via env files or vault agents instead of baking them into images.
1 doc page
- →
Documentation
Guides, tutorials, references, and architecture notes explain how to build and operate on the platform safely. Searchable portals pair human prose with OpenAPI and Terraform schemas assistants can index. Each page states prerequisites, defaults, and failure modes to shorten incident time.
1 doc page
- →
Dvr
Distributed virtual routing spreads Layer-3 and NAT functions across compute nodes instead of central network nodes. DVR pairs with floating IPs and tenant subnets for scalable east-west and north-south traffic. Troubleshoot symmetric paths when pings succeed one direction only.
1 doc page
- →
Dynamic Inventory
1 doc page
- →
Ecommerce
2 doc pages
- →
Egress
Outbound traffic leaves project networks toward the internet, VPN peers, or other VPCs and often meters separately. Spikes may signal exfiltration, misconfigured mirrors, or backup jobs. Security groups and routes must allow return paths for established flows.
1 doc page
- →
Embeddings
Dense vectors represent text, images, or code chunks so nearest-neighbor search ranks semantic similarity. Embedding models run at index time and sometimes query time inside RAG stacks. Normalization and distance metrics affect recall and latency trade-offs.
2 doc pages
- →
Encryption
Algorithms transform plaintext into ciphertext using keys managed by KMS, disk subsystems, or TLS sessions. Envelope encryption limits exposure when a single data key leaks. Regulations often mandate algorithms, key lengths, and rotation evidence.
4 doc pages
- →
Endpoints
Service catalog entries list public, internal, and admin URLs clients use to reach each OpenStack API surface. Misconfigured endpoints cause auth loops, wrong-region calls, or certificate hostname errors. Automation should pick the interface matching network reachability.
1 doc page
- →
Error Handling
APIs return actionable HTTP status codes, correlation IDs, and typed problem details clients can branch on. Exponential backoff with Retry-After headers prevents thundering herds during incidents. Idempotency keys stop duplicate charges when clients safely retry writes.
11 doc pages
- →
Filesystems
Layouts such as ext4 or XFS organize blocks into files, directories, permissions, and journals on attached disks. Cloud images ship with tuned mount options while data volumes format independently. Object storage avoids POSIX yet still serves key-value access patterns.
1 doc page
- →
Flavors
Named virtual hardware templates specify vCPU count, RAM, root disk class, and optional extras like huge pages. Curated flavors keep capacity planning predictable for finance and scheduling teams. Resizing may require shutdown when disk footprints shrink.
28 doc pages
- →
Floating IPs
Public addresses drawn from provider pools attach to instance ports for stable inbound connectivity. They decouple client-facing IPs from migrations or rebuilds behind the scenes. DNS A records pointing at floats simplify failover drills.
143 doc pages
- →
Flowise
Visual node editors wire LLM prompts, retrievers, and HTTP tools into shareable agent graphs. Teams prototype assistants without boilerplate then export JSON for self-hosted runtimes. Inject secrets via environment variables and audit outbound calls like any automation.
1 doc page
- →
Gateways
Logical next hops connect tenant subnets to external providers, VPN concentrators, or transit fabrics. They apply default routes, SNAT, and sometimes BGP advertisements. Broken gateway config yields asymmetric routing or hairpinned traffic.
1 doc page
- →
Glance
OpenStack Image service stores bootable disk templates, metadata, and visibility across projects. Nova pulls images when launching instances while operators upload qcow2 or raw formats. Backing object storage scales out large image blobs with checksum verification.
2 doc pages
- →
Glossary
Alphabetized definitions anchor terminology for writers, translators, and support engineers so everyone uses the same words for the same things. Cross-links reduce duplicated explanations inside long guides while autolinks connect prose to canonical entries automatically. Regular reviews retire deprecated marketing names and capture new product language before it spreads inconsistently.
1 doc page
- →
Hardening
Teams shrink attack surface via patching, CIS baselines, disabling unused services, and enforcing MFA plus TLS 1.2+. Image pipelines rebuild on CVE notices while seccomp profiles constrain containers. Evidence packets satisfy auditors faster than ad hoc screenshots.
1 doc page
- →
Heat
OpenStack native orchestration using HOT (YAML) templates that declare networks, instances, security, and supporting resources as one stack. Legacy on Quake AI — use OpenTofu for new infrastructure automation. Stack events timeline failures while updates stay idempotent when templates drift.
11 doc pages
- →
High Availability
Architectures combine redundant components, health checks, and failover to cap downtime within SLO budgets. Multi-AZ placement, load balancing, and database replication are common patterns. Untested failover exercises hide brittle DNS or session assumptions.
4 doc pages
- →
Hot Templates
1 doc page
- →
IaC
Git-reviewed specifications describe desired infrastructure so planners and appliers recreate environments repeatably. OpenTofu or Terraform plans pair with policy checks before merge. Heat templates are a legacy alternative. Treat infra changes like application releases with semver and rollback plans.
35 doc pages
- →
Iam
Identity and access management on AWS, Azure, and GCP covers users, roles, policies, and service principals. Quake AI scopes access through Rumble.com accounts, organizations, projects, and Keystone roles instead of per-resource IAM policies.
1 doc page
- →
Identity
Keystone manages users, groups, projects, domains, roles, and application credentials across the cloud. Federation maps corporate IdP attributes to cloud roles for single sign-on. Service users authenticate automation separately from humans.
5 doc pages
- →
Images
Bootable templates bundle OS bits, drivers, and disk layout metadata consumed at instance launch. Metadata captures minimum disk, boot modes, and hardware property hints. Golden images rebuild through CI whenever security baselines change.
39 doc pages
- →
Incident Response
Structured workflows detect breaches or outages, contain blast radius, eradicate root cause, and communicate timelines. Runbooks assign roles, escalation paths, and evidence preservation steps. Postmortems feed automation and training for the next event.
1 doc page
- →
Inference
Serving stacks run forward passes through trained models to answer prompts, score embeddings, or classify inputs. Autoscaling watches queue depth while quantization trades accuracy for latency. Metrics on tokens per second and error rates guide capacity planning.
3 doc pages
- →
Infrastructure As Code
Treating servers, networks, and policies as versioned artifacts enables reproducible applies and drift detection. Pipelines lint, plan, and apply changes with human approvals on risky resources. Modules capture org standards the same way shared libraries do for apps.
2 doc pages
- →
Ingress
North-south client traffic enters reverse proxies, routers, or Kubernetes clusters toward internal services. Operators configure listeners, certificates, and ACLs at those edges. Symmetric egress rules must exist for return flows through complex topologies.
1 doc page
- →
Instances
Running virtual machines combine hypervisor CPU, RAM, ports, volumes, and metadata scheduled from images and flavors. Cloud-init finishes provisioning while APIs expose power states and serial consoles. Autoscaling groups grow or shrink counts based on policies you define.
241 doc pages
- →
Invoices
2 doc pages
- →
Ip Addresses
IPv4 or IPv6 identifiers locate interfaces on routed networks with static allocation or DHCP pools per subnet. Planning avoids overlapping CIDRs across peered VPCs and on-prem links. DNS and floating IPs abstract clients from underlying instance churn.
3 doc pages
- →
Key Pairs
SSH public keys registered in the cloud inject into authorized_keys during boot while private keys stay on workstations. Rotate by uploading new public material and retiring old logins. Passphrase-protected private keys limit impact if laptops leak.
25 doc pages
- →
Keystone
OpenStack identity service issues scoped tokens, publishes the service catalog, and enforces RBAC across APIs. Other components trust Keystone validation and revocation lists. Federation integrates corporate IdPs without duplicating user stores.
3 doc pages
- →
Kubernetes
Open-source control plane schedules pods, services, ingress, storage, and rolling updates across worker nodes. Declarative manifests reconcile desired state continuously. CNI plugins rely on solid underlay networking from the IaaS layer.
50 doc pages
- →
Langchain
Framework modules chain prompts, parsers, retrievers, and tool executors into composable LLM applications. Swap models or vector stores without rewriting orchestration glue. Tracing hooks expose per-step latency for tuning agent loops.
1 doc page
- →
Launch
7 doc pages
- →
Legal
1 doc page
- →
Lifecycle
1 doc page
- →
Limits
Per-project ceilings on cores, instances, floating IPs, volumes, and API rates protect shared capacity. Operators raise limits after capacity reviews while automation surfaces friendly quota errors. Dashboards trending toward limits prevent launch-day surprises.
1 doc page
- →
LLM
Large language models predict next tokens from prompts, tool outputs, and retrieved context windows. Fine-tuning steers tone while guardrails reduce unsafe completions. Cost, latency, and memory dominate hosting decisions at scale.
5 doc pages
- →
Load Balancer
A self-managed reverse proxy distributes traffic across backends using routing and health-check rules. Kubernetes workloads can request a public endpoint with a Service of type LoadBalancer. Configure TLS termination on the reverse proxy, ingress controller, or application.
1 doc page
- →
Logging
Timestamped streams from apps, hypervisors, APIs, and OS facilities land in centralized stores for search and retention. Structured JSON fields accelerate parsing while correlation IDs stitch distributed traces. Alerts fire on error spikes or missing heartbeats from critical jobs.
2 doc pages
- →
Mac Addresses
Layer-2 identifiers on Ethernet interfaces identify frames inside broadcast domains before routing takes over. OpenStack ports receive generated or custom MAC values that must stay unique per segment. Security inventories track MACs for asset management and anomaly detection.
1 doc page
- →
Magnum
OpenStack API for container orchestration engines such as Kubernetes. On Quake AI, Magnum provisions Kubernetes clusters from cluster templates; provisioning self-managed Kubernetes on Nova instances with OpenTofu is the other supported path, and both remain available. Cluster templates encode COE versions, networking, and autoscaling policies.
4 doc pages
- →
Managed Vs Self Hosted
Vendors operate upgrades, backups, and patching in managed offerings while self-hosted stacks maximize control and data residency. Hybrid mixes offload commodity layers but keep regulated data on owned hardware. Total cost includes engineer time, not only line-item invoices.
1 doc page
- →
Mcp
Model Context Protocol lets assistants discover tools, prompts, and resources over stdio or streamable HTTP transports. Typed schemas prevent surprise shell execution when clients enforce approvals. Servers wrap internal APIs for desktop agents without exposing raw admin networks.
4 doc pages
- →
Microversions
OpenStack services add optional request and response fields through incremental API sub-versions without major bumps. Clients advertise supported ranges via headers while servers publish min and max compatibility. Pin automation to explicit microversions to avoid surprise field drift during upgrades.
1 doc page
- →
Migration
Workloads move across hypervisors, regions, or providers using live or cold instance migration plus storage sync. Runbooks coordinate DNS cuts, reverse-proxy drains, and database ordering to minimize downtime. Validate networking and identity before declaring migration complete.
71 doc pages
- →
Model Quantization
Reducing weight precision from FP16 or FP32 toward INT8 or INT4 shrinks VRAM footprints and speeds matmuls. Accuracy shifts per task, so benchmark perplexity or task metrics before production. Serving stacks expose flags to pick precision per deployment tier.
1 doc page
- →
Monitoring
Collectors scrape metrics, tail logs, and evaluate SLO burn rates to page owners before users notice regressions. Dashboards pair application golden signals with infrastructure saturation charts. Synthetic checks validate critical journeys even when traffic is quiet.
6 doc pages
- →
Multi Cloud
Architectures deliberately span providers or regions for resilience, data residency, or commercial leverage. Portable images, Terraform modules, and container images reduce lock-in friction. Harmonized tagging and cost exports keep finance sane across estates.
1 doc page
- →
Multi Environment
1 doc page
- →
n8n
Open-source workflow automation connects webhooks, HTTP nodes, schedules, and credentials through a visual editor. Self-hosting keeps execution data inside your perimeter while community nodes expand integrations. Treat runs like any batch job with retries, idempotency, and secret hygiene.
1 doc page
- →
NAT
Address rewriting lets many private hosts share fewer public IPv4 addresses through connection state tables. SNAT handles outbound flows while DNAT publishes inbound services. Routers and floating IP edges enforce provider NAT policies.
2 doc pages
- →
Navigation
1 doc page
- →
Networking
VPCs, subnets, routers, security groups, DNS, VPNs, and floating IPs stitch workloads together with audited paths. Operators segment environments for compliance while CNIs consume the underlay from Kubernetes. MTU and asymmetric routing tests catch subtle breakage.
20 doc pages
- →
Networks
Isolated Layer-2 segments attach instance ports, supply DHCP, and host metadata agents before routing begins. Provider networks map to datacenter VLANs while tenant overlays use encapsulation. Plan address space to avoid collisions with peered links.
155 doc pages
- →
Nginx
13 doc pages
- →
Object Storage
RESTful buckets store immutable blobs keyed by names with rich metadata, versioning, and lifecycle rules. Erasure coding and replication deliver durability without POSIX semantics. Clients use HTTP verbs, SDKs, or pre-signed URLs for large-scale media and backups.
83 doc pages
- →
Observability
Metrics, logs, traces, and profiles explain internal state from externally observable signals without shell access. OpenTelemetry standardizes instrumentation while cardinality discipline controls backend cost. SLOs tie observability budgets to customer-facing reliability targets.
3 doc pages
- →
Ollama
Local runtime downloads GGUF weights and serves chat or embedding HTTP APIs with simple CLI ergonomics. GPU acceleration depends on bundled drivers matching your hardware. Air-gapped sites mirror models and verify checksums before promotion.
4 doc pages
- →
Onboarding
2 doc pages
- →
Open Webui
Browser chat UI fronts local or remote LLM backends with model switching, plugins, and lightweight RBAC. Self-hosting preserves data residency for private ChatGPT-like experiences. Pair with VPN or SSO when exposing it beyond localhost.
1 doc page
- →
Openclaw
Open-source AI agent framework connects frontier or local models to messaging platforms like Telegram, Slack, and Discord. Operators deploy gateway containers on VMs, store keys under home directories, and expose webhooks via floating IPs. Heat templates can stand up instances, security groups, and addresses automatically.
1 doc page
- →
OpenStack
Open-source cloud operating system projects deliver compute, storage, networking, identity, and orchestration APIs. Nova, Neutron, Cinder, Keystone, Glance, and Heat form common building blocks operators integrate. Quake AI exposes those primitives through portals and automation tailored for developers.
6 doc pages
- →
OpenStack CLI
Unified openstack commands wrap project APIs for servers, networks, images, stacks, and secrets using shared auth plugins. Environment variables or clouds.yaml select clouds, regions, and interface URLs. Script-friendly output formats integrate with jq and CI systems.
21 doc pages
- →
Opentofu
Community-governed fork compatible with Terraform HCL plans and applies infrastructure changes with remote state locking. Providers target OpenStack and many other APIs from the same workflow. Validation gates in CI catch unsafe diffs before merge.
5 doc pages
- →
Operations
Day-two work covers patching, scaling, incident response, capacity planning, and customer communications after launch. Runbooks encode on-call expectations while metrics show when to order hardware. Blending automation with human judgment keeps SLOs honest.
15 doc pages
- →
Orchestration
1 doc page
- →
Organizations
Plural usage highlights multiple enterprise containers each with distinct members, policies, and spend caps. Consolidated billing can span several organizations only when contracts explicitly allow. Audit trails should attribute actions to both org and project scopes.
4 doc pages
- →
Payment Methods
1 doc page
- →
Placement Policies
Scheduler hints, aggregates, availability zones, and PCI passthrough requests steer workloads to capable hosts. Flavor extra specs encode hardware needs Nova evaluates against inventory. Placement failures mean constraints exceed free capacity rather than generic errors.
2 doc pages
- →
Platform
4 doc pages
- →
Policy
1 doc page
- →
Portal
Web experience unifies documentation search, support tickets, billing, and deployment wizards behind single sign-on. Status embeds and changelogs reduce tab hopping during incidents. Deep links carry auth context into consoles safely.
3 doc pages
- →
Ports
Neutron ports bind MAC addresses, fixed IPs, security groups, and attachments to instances or routers. Floating IPs associate with ports for inbound services while status fields show binding health. Automation can create ports before attaching them to instances or routers.
12 doc pages
- →
Pricing
Published rates for flavors, storage tiers, egress, and support SKUs anchor forecasts and procurement conversations. Dynamic calculators estimate monthly totals from expected usage patterns. Discounts and commits still need engineering visibility to avoid surprise bills.
4 doc pages
- →
Projects
Isolated namespaces own quotas, networks, instances, roles, and object containers separate from sibling projects. Keystone scopes tokens to a project so APIs cannot cross boundaries accidentally. Many enterprises map projects to environments like staging and production.
10 doc pages
- →
Public Key Cryptography
Asymmetric math lets anyone encrypt or verify with a public key while only holders of the private key decrypt or sign. SSH, TLS, and X.509 chains rely on this split to bootstrap trust. Plan migrations toward post-quantum algorithms as standards mature.
1 doc page
- →
Qcow2
QEMU copy-on-write image format stores sparse disks with optional backing files and snapshot chains. Glance often distributes qcow2 templates while Cinder may flatten to raw for predictable performance. Thin provisioning saves space until writes allocate real blocks.
2 doc pages
- →
Qdrant
Open-source vector database indexes embeddings for similarity, hybrid filters, and metadata-rich retrieval workloads. Teams self-host collections beside private documents for semantic search and RAG. Tuning HNSW parameters balances recall against query latency.
2 doc pages
- →
Quake Yaml
6 doc pages
- →
Quotas
Numeric caps per project limit instances, cores, RAM, volumes, floating IPs, and API rates to protect shared fleets. Dashboards trending toward limits trigger quota increase requests before launches fail. Automation should parse quota errors distinctly from permission denials.
4 doc pages
- →
Rag
Retrieval-augmented generation grounds model answers with chunks from vector or keyword indexes before decoding tokens. Chunking, hybrid search, and rerankers control relevance while citations reduce hallucinations. Refresh embeddings whenever documentation meaningfully changes.
4 doc pages
- →
Ram
Volatile memory attached to instances via flavors must satisfy working-set needs without constant swap thrash. JVM heaps, databases, and caches drive RAM sizing more than CPU counts alone. Hot-plug support varies by platform, so plan maintenance windows accordingly.
2 doc pages
- →
Rate Limits
1 doc page
- →
Rclone
Command-line sync utility copies objects between S3-compatible endpoints, Swift, POSIX, and dozens of remotes with checksums. Bandwidth limits and retries make it ideal for petabyte migrations or CI artifact publishing. Encrypt remotes when crossing trust boundaries.
4 doc pages
- →
Regions
Discrete datacenter footprints expose separate API endpoints, fault domains, and capacity pools. Latency, data residency, and DR pairing drive region selection. Cross-region replication adds egress and storage costs planners must model.
2 doc pages
- →
Regions And Zones
Hyperscalers partition capacity into regions and availability zones for fault isolation and latency. Quake AI runs in named regions; placement and redundancy follow project networking and instance design rather than a separate AZ SKU list.
1 doc page
- →
Resilience
2 doc pages
- →
Resource Tiers
Classes such as burstable versus guaranteed CPU or standard versus performance volumes price predictably for different SLAs. Architects map tiers to workload criticality while autoscaling policies may shift tiers under load. Documentation ties tier names to underlying hardware profiles.
5 doc pages
- →
Retry
2 doc pages
- →
Roles
2 doc pages
- →
Root Disk
Primary boot volume hosts the OS, packages, and often application state unless you attach separate data disks. Ephemeral root disks vanish on delete while volume-backed roots persist independently. Resize workflows must grow filesystems safely after flavor changes.
1 doc page
- →
Routers
Neutron routers connect tenant subnets to external gateways, apply default routes, and enable SNAT for outbound flows. DVR mode distributes forwarding across compute nodes for scale. Static routes send traffic toward VPN appliances or inspection tiers.
22 doc pages
- →
Runbooks
Step-by-step procedures cover incidents, upgrades, migrations, and rollback validation with commands and dashboards linked. Checklists help newcomers while decision trees guide veterans during ambiguity. Post-incident updates capture lessons for the next rotation.
2 doc pages
- →
S3
Amazon's object API model inspires compatible endpoints listing buckets, uploading objects, and using multipart uploads. Rumble exposes S3-compatible URLs on Swift-backed storage with documented dialect differences. SDKs point custom endpoints and signing regions at the provider configuration.
31 doc pages
- →
S3 Api
HTTP surface implements bucket operations, object CRUD, multipart uploads, and conditional headers familiar to AWS SDKs. Signature versions and endpoint styles must match what the provider supports. Integration tests validate error codes your automation relies on.
2 doc pages
- →
S3 Compatibility
Measures how closely an endpoint matches AWS semantics for ACLs, versioning, encryption headers, and error shapes. Gaps may require client-side workarounds or adjusted concurrency assumptions. Publish matrices so migrations from hyperscalers proceed with eyes open.
2 doc pages
- →
S3 Credentials
Access keys, secrets, and session tokens authenticate SDKs against S3-compatible endpoints with scoped policies. Rotate leaked pairs immediately and never store them in git. Prefer STS-style temporary credentials for automation when available.
3 doc pages
- →
Secrets
Confidential material includes API tokens, TLS private keys, database passwords, and signing seeds requiring vault storage. Rotation policies and sealed CI integrations limit exposure windows. Logs must redact secret values even during verbose debugging.
5 doc pages
- →
Security
Holistic programs blend identity hardening, vulnerability management, encryption, detective controls, and secure SDLC practices. Shared responsibility charts clarify who patches guests versus hypervisors. Threat modeling informs network segmentation and least-privilege defaults.
25 doc pages
- →
Security Groups
Stateful distributed firewall rules attach to Neutron ports to allow or deny ingress and egress tuples explicitly. Default-deny posture plus narrow allows beats giant CIDR lists nobody audits. Referencing remote groups scales rules across tiers of instances.
133 doc pages
- →
Self Hosted Ai
Running models, vector databases, and agent gateways on owned hardware preserves data residency and cuts per-token SaaS bills. Operators still patch GPUs, watch capacity, and tune batching like any other service. Air-gapped environments depend on this pattern exclusively.
6 doc pages
- →
Server Groups
Nova constructs apply affinity, anti-affinity, or soft variants to steer instance placement across hypervisors. Licensing or latency-sensitive tiers may require co-location while resilient tiers spread out. Policies attach at launch and interact with scheduler filters.
10 doc pages
- →
Server Side Encryption
1 doc page
- →
Servers
Generic cloud language and OpenStack APIs use server for a running virtual machine. Quake AI product documentation prefers instance while Nova commands such as openstack server list still use server.
1 doc page
- →
Services
OpenStack service catalog entries advertise APIs for compute, identity, image, volume, network, orchestration, and more. Each long-running daemon upgrades on rolling schedules with health checks. Application teams also call their microservices services—context disambiguates the layer.
2 doc pages
- →
Shared Responsibility
Contracts split obligations so providers secure facilities and hypervisors while customers harden guests, data, and IAM. Misunderstood lines cause audit gaps after incidents. Documentation maps each control to the responsible party clearly.
2 doc pages
- →
Sla
1 doc page
- →
Snapshot Types
Providers distinguish instance or VM snapshots from volume or disk snapshots with different APIs and restore paths. Quake AI documents image-backed instance capture separately from Cinder volume snapshots and clones.
1 doc page
- →
Snapshots
Point-in-time volume or image captures enable rollback, cloning, and incremental backup chains. Application-consistent hooks quiesce databases when supported before freezing blocks. Retention policies balance recovery needs against storage growth costs.
27 doc pages
- →
Snat
Source NAT rewrites private addresses to shared public IPs for outbound internet access through tenant routers. Stateful tables ensure return packets map back to the correct instances. Default SNAT on gateways complements explicit floating IPs for inbound services.
1 doc page
- →
SSH
Encrypted interactive shells, file copies, and tunnels traverse TCP port 22 with host keys pinned in known_hosts files. Key-based authentication replaces password logins for automation and humans alike. Jump hosts aggregate access into private subnets without exposing every instance publicly.
14 doc pages
- →
Ssh Keys
Public key material you register for Linux login is often called an SSH key in tutorials and other clouds. Quake AI stores the same material as key pairs injected through cloud-init at instance boot.
2 doc pages
- →
Stacks
Heat-managed collections of resources create, update, or delete together while events log failures for triage. Part of the legacy Heat orchestration service. For new work, use OpenTofu to manage resource groups. Outputs export URLs and IDs for dependent pipelines.
1 doc page
- →
State Management
1 doc page
- →
Stateful Filtering
Firewalls track connection state so related packets flow without duplicating explicit rules for every phase. OpenStack security groups are stateful by default, simplifying allow lists. Large connection tables consume memory; tune timeouts for long-lived protocols.
1 doc page
- →
Storage
Umbrella term spans block volumes, object buckets, file shares, snapshots, replication, and encryption policies. Architects pick modalities based on latency, consistency, durability, and cost. Capacity monitoring prevents emergency expansions during peak events.
19 doc pages
- →
Subnets
CIDR slices inside a network allocate gateway IPs, DHCP pools, DNS options, and availability-zone affinity. Routers join subnets for east-west routing while planning avoids overlap with peered networks. IPv6 dual-stack designs mirror IPv4 policies for parity.
15 doc pages
- →
Subscriptions
Multiple active subscriptions can layer production entitlements with sandbox add-ons across one organization without merging their billing narratives. Finance reconciles each subscription line against usage exports and contract schedules every close. Documentation should clarify which APIs, quotas, and support SLAs each tier unlocks so engineers provision with correct expectations.
8 doc pages
- →
Support
Provider teams answer tickets, run incident bridges, and supply root-cause analyses within SLA windows. Severity definitions align customer expectations with response clocks. Self-serve status pages complement human support during widespread events.
3 doc pages
- →
Swift
OpenStack object storage project delivers durable, eventually consistent buckets with erasure coding and replication. S3-compatible gateways often sit in front of Swift for SDK interoperability. Operators scale out storage nodes horizontally as capacity grows.
2 doc pages
- →
Team Roles
RBAC assignments such as member, reader, or billing admin map people to permitted actions inside projects and orgs. Least privilege limits blast radius when credentials leak. Periodic access reviews revoke stale grants after reorganizations.
1 doc page
- →
Templates
1 doc page
- →
Terraform
HashiCorp tool and ecosystem apply declarative HCL plans across clouds with remote state and policy integrations. OpenStack provider resources mirror APIs for networks, instances, and identities. Many teams standardize on Terraform modules for baseline architectures.
52 doc pages
- →
Tls
Transport Layer Security encrypts HTTP and other protocols while authenticating servers via X.509 chains. Modern configs prefer TLS 1.2 or 1.3 with strong cipher suites and HSTS where browsers apply. Certificate automation via ACME reduces manual renewal toil.
3 doc pages
- →
Tokens
Keystone issues scoped bearer credentials clients place in headers to authorize API calls until expiry or revocation. Services validate tokens centrally and honor role assignments embedded in payloads. Short lifetimes plus refresh flows limit stolen-token damage.
2 doc pages
- →
Transfers
1 doc page
- →
Troubleshooting
Structured diagnosis gathers symptoms, reproduces minimally, checks recent changes, and validates dependencies end-to-end. Correlation IDs and packet captures narrow whether faults sit in app, network, or identity layers. Document fixes back into runbooks when novel root causes appear.
9 doc pages
- →
Tutorials
Hands-on walkthroughs teach a skill by completing a realistic task with checkpoints and expected outputs. They pair narrative with code blocks users can paste into terminals. Link to reference docs when parameters multiply beyond tutorial scope.
8 doc pages
- →
Validation
1 doc page
- →
Vcpu
Virtual CPU threads map to hypervisor pCPUs with optional oversubscription ratios affecting noisy-neighbor risk. Pinning and NUMA topology hints target latency-sensitive workloads. Right-sizing flavors avoids paying for idle compute capacity.
2 doc pages
- →
Vector Search
Query pipelines embed user text, retrieve nearest neighbors, optionally rerank, then pass chunks to downstream models. Tuning top-k, distance metrics, and score thresholds balances recall and precision. Fresh indexes matter whenever corpora change materially.
3 doc pages
- →
Versioning
Object stores and APIs track multiple generations of the same key or contract so readers pick explicit revisions. Lifecycle rules expire old versions to control storage costs. Clients must enable versioned buckets when rollback semantics matter.
5 doc pages
- →
Virtual Machines
Industry-generic VM language describes an isolated guest OS with virtual CPU, memory, and disks. On Quake AI that entity is an instance in the Compute service with the same lifecycle and APIs.
3 doc pages
- →
Volumes
Cinder block devices attach to instances for persistent data beyond ephemeral root disks. Snapshots and clones accelerate database refreshes and DR testing. Performance tiers map to backend pools with replicated or local flash characteristics.
105 doc pages
- →
VPN
Encrypted tunnels connect on-prem networks or laptops to cloud VPCs over the public internet or dedicated links. Route-based versus policy-based VPNs influence how subnets advertise reachability. Monitor tunnel health and MTU issues after firmware upgrades on either side.
3 doc pages