Skip to content
Claresia Docs

Status page interpretation

status.claresia.com is the public, real-time view of Claresia service health. Designed in the lineage of status.stripe.com. Subscribe by region + service so you only see pings relevant to you.

Layout

Section titled “Layout”

The page has three rows of components and an incident timeline below:

  1. Region row — one per region (eu-south-1 (Milano), eu-central-1 (Frankfurt), …)
  2. Service row — one per service inside each region
  3. Component cells — green / yellow / red

Services tracked

Section titled “Services tracked”

Each region has these services:

ServiceWhat it covers
Identity (WorkOS)SSO sign-in, SCIM sync, JWT issuance
Command CenterAdmin console, Hub Browser, Settings UI
Hub (Postgres)Read + write + provenance verification
Distribution Plane — AnthropicSkill publish to Claude Enterprise
Distribution Plane — MicrosoftSkill publish to Copilot M365
Distribution Plane — OpenAISkill publish to ChatGPT Enterprise
Distribution Plane — SlackSkill publish to Slack workspaces
Telemetry PipelinePull connectors + ingest + Command Center surface
Onboarding PortalNew-tenant + existing-tenant config flows
Documentation SiteThis site

What each color means

Section titled “What each color means”
  • 🟢 Green — Service operating within SLO. No customer-impacting issues.
  • 🟡 Yellow — Degraded. Some users may experience slowness or partial failure. SLO trending toward burn.
  • 🔴 Red — Outage. SLO violated. Active P1 incident in flight.
  • ⚪ Maintenance — Planned maintenance. Posted at least 14 days in advance via email + Trust Center.
  • ⚙️ Operational with notice — Service is up but a notable change is in flight (e.g., a major skill IR version cutover).

How to read combinations

Section titled “How to read combinations”
What you seeWhat it means
All green in your regionEverything’s fine. No action needed.
Yellow on one Distribution PlaneSkill publishes to that LLM platform are slower than 60s SLO; existing publishes still work. End users are not impacted unless you’re actively republishing.
Yellow on Telemetry PipelineCommand Center stats may be stale up to 30 min. End-user experience unaffected.
Red on HubSkill invocations cannot write outputs. End users see skill failures. Sev 1 — declare in your channel.
Red on IdentitySSO sign-in fails for all users. Sev 1 — declare immediately.
Yellow on Identity in your region onlyLikely an IdP-side issue; check your IdP’s status page first.
Red on Documentation SiteThis site is down. Doesn’t affect your operations; just inconvenient.

Subscribing

Section titled “Subscribing”

Visit status.claresia.com/subscribe:

  • Email — daily digest + immediate Sev 1
  • Slack webhook — to your Slack workspace, per-channel
  • RSS/Atom — for your feed reader
  • Generic webhook — for your monitoring system (PagerDuty, OpsGenie, etc.)

Subscribe at the service level (only Hub, only Identity, etc.) and the region level so you only see pings relevant to you.

Incident timeline

Section titled “Incident timeline”

Below the component grid:

  • Active incidents — currently red or yellow, top of page
  • Past 30 days — chronological list of resolved incidents with PIR links
  • Past 90 days — uptime % per region per service (the historical compliance evidence)

Each incident card shows:

  • Title (one-line)
  • Affected services + regions
  • Started + resolved timestamps
  • Severity (Sev 1 / 2 / 3)
  • Latest update message
  • PIR link (post-incident review, available within 5 business days of resolution)

Maintenance schedule

Section titled “Maintenance schedule”

A rolling 14-day maintenance schedule is published:

  • Maintenance windows are typically Sunday 02:00–04:00 UTC for the affected region
  • Posted at least 14 days in advance via:
    • Status page banner
    • Email to it-admin-contacts set in Command Center
    • In-app banner in Command Center

For breaking changes, 90 days notice plus a feature-flag rollback path.

Historical uptime

Section titled “Historical uptime”

Per region per service, last 90 days:

Region        Service                 Uptime   Incidents
eu-south-1     Identity (WorkOS)       99.99%   0
eu-south-1     Command Center          99.97%   1 (sev3)
eu-south-1     Hub                     99.95%   1 (sev2)
eu-south-1     Distribution-Anthropic  99.91%   2 (sev2, sev3)
eu-south-1     Telemetry Pipeline      99.80%   3 (sev2)
eu-south-1     Onboarding Portal       99.99%   0
eu-south-1     Documentation Site     100.00%   0
eu-central-1  ...                     ...      ...

Historical uptime is the public-evidence artifact for SLA claims. Mode B customers have an SLA contract that references the same numbers.

What you should NOT use the status page for

Section titled “What you should NOT use the status page for”
  • Real-time end-user issue diagnosis — start with your tenant’s Command Center health view (faster, more granular)
  • Forecasting LLM costs — use the Telemetry tab in Command Center
  • Sub-incident-level triage — request the PIR after resolution

SLO source-of-truth

Section titled “SLO source-of-truth”

The status page is automatically driven by Claresia’s internal SLO burn-rate alerts. We don’t manually update it during an incident — it reflects the actual measured state of every service. Our SRE team is on-call against the same signal you see.

Maintenance windows by region

Section titled “Maintenance windows by region”
RegionStandard window
eu-south-1Sunday 02:00–04:00 EST (07:00–09:00 UTC)
eu-central-1Sunday 02:00–04:00 CET (01:00–03:00 UTC)
ap-southeast-1 (planned)Sunday 02:00–04:00 SGT (18:00–20:00 UTC Sat)

Maintenance only affects services that explicitly require downtime (rare — most updates are zero-downtime via blue-green deploys).