Platform status

• Baz CloudBaz itself: the app, the API and the review workers.

  All services are online99.69%
  90 days agoToday
  Show 5 components
• GitHubWhere most Baz reviews are triggered: pull requests, webhooks and Actions.

  All Systems Operational95.05%
  90 days agoToday
  Show 11 components
• GitLabRepository hosting, merge requests and CI for teams on GitLab.

  Operational98.57%
  90 days agoToday
  Show 23 components
• Azure DevOpsAzure Repos, Pipelines and Boards for teams on Microsoft's stack.

  Everything is looking good96.44%
  90 days agoToday
  Show 7 components
• OpenAIModels behind Codex and the OpenAI-backed review paths.

  All Systems Operational96.17%
  90 days agoToday
  Show 25 components
• AnthropicClaude models behind Baz reviews and Claude Code.

  All Systems Operational96.99%
  90 days agoToday
  Show 6 components
• AWS BedrockManaged model inference behind Baz reviews running on AWS.

  No open AWS events affecting Bedrock100%
  90 days agoToday
  Show 6 components

• MinorOpenAISep 8, 2:39 PM UTC4.6 h

  File uploads are delayed or failing

  All impacted services have now fully recovered.

• MinorOpenAISep 8, 2:32 PM UTC7.4 h

  Elevated errors for image generation

  All impacted services have now fully recovered.

• MinorGitLabSep 8, 8:54 AM UTC1.1 h

  Some catalog agents are not working

  Resolved - The merge request that introduced the failure was reverted, and custom/catalog agents are working again. More information is available in gitlab.com/gitlab-com/gl-infra/production/-/work_items/22880.

• MinorAzure DevOpsSep 7, 2:14 PM UTC17 h

  Availability Degradation - Pipelines, MDP pools provisioning (Asia Pacific)

  <p>The issue is now fully mitigated. Our engineers will be investigating this further to learn from and reduce the risk of potential recurrences. We apologize for the impact this had on our customers.</p>

• MinorGitHubSep 4, 10:02 PM UTC21 min

  Degradation in repos contents API

  This incident has been resolved. Thank you for your patience and understanding as we addressed this issue. A detailed root cause analysis will be shared as soon as it is available.

• MajorGitHubSep 4, 8:39 PM UTC1.8 h

  Disruption with Copilot Code Review

  This incident has been resolved. Thank you for your patience and understanding as we addressed this issue. A detailed root cause analysis will be shared as soon as it is available.

• Baz Cloud

  Baz's own Better Stack status page: live resource states and published uptime read from the page, incident windows from its RSS feed grouped by incident.

• GitHub

  Atlassian Statuspage v2 API: live component states, history from the published incident feed and its tagged impact.

• GitLab

  Status.io public status API for live component states, plus the page's incident RSS feed. Severity is inferred from the incident text, since Status.io does not publish a machine-readable impact per incident.

• Azure DevOps

  Azure DevOps health API for live per-service and per-geography health, plus the recent events RSS feed. Incident windows come from the timestamps Microsoft prints in each update, and severity is inferred from the event title.

• OpenAI

  Atlassian Statuspage v2 API: live component states, history from the published incident feed and its tagged impact.

• Anthropic

  Atlassian Statuspage v2 API: live component states, history from the published incident feed and its tagged impact.

• AWS Bedrock

  AWS Health public events feed for anything currently affecting Bedrock, plus the per-region Service Health Dashboard RSS feeds. Consecutive regional updates within six hours are merged into a single incident window.

Daily status is reconstructed from published incident windows: a day takes the worst impact of anything overlapping it, and uptime is one minus the union of the minor and major windows over the period. Announced maintenance colours a day but does not count against uptime. Providers that do not publish a machine-readable severity have theirs inferred from the incident text, so treat those rows as indicative. Refreshed every two minutes; the raw aggregate is available at /api/status.