Integrations FAQ
Which alert sources are supported?
OpsWorker supports two alert sources: Prometheus AlertManager and Datadog. Both connect via webhook. Grafana is not a native alert source, but Grafana Alerting (9+) works because it sends alerts in AlertManager-compatible format to the AlertManager webhook endpoint.
Can I use multiple alert sources for one cluster?
Yes. You can connect AlertManager (including Grafana Alerting in AlertManager format) and Datadog to the same cluster simultaneously. OpsWorker normalizes alerts from all sources into a common format.
Does OpsWorker replace my monitoring tools?
No. OpsWorker is an investigation layer that works on top of your existing monitoring. Your Prometheus, Grafana, and Datadog setups remain unchanged. OpsWorker investigates the alerts they fire.
How does the Slack integration work?
OpsWorker connects to Slack via OAuth. After authorization, investigation results are posted to your configured channels with root cause, recommendations, and feedback buttons. See Slack Integration.
What does the GitHub/GitLab integration provide?
It enables the Source Code & Repository Agent, which correlates code changes with infrastructure incidents. During investigations and chat, the agent can view commits and pull requests to identify whether a recent change caused an issue. The integration requires a connected GitHub or GitLab account.
Can I use OpsWorker without Slack?
Yes. All investigation results are available in the OpsWorker portal. Slack is optional but recommended; it delivers results where your team already works.
What AI models does OpsWorker use?
OpsWorker uses AWS Bedrock (Amazon Nova for routing and summarization, Claude for reasoning) and the direct Anthropic API. Nova handles routing and summarization, not field extraction; alert field extraction is deterministic and regex-based, not LLM-driven. You don't need to configure anything for the default setup.
Is Grafana MCP different from Grafana Alerting?
Yes. Grafana Alerting sends alerts from Grafana to OpsWorker for investigation. Grafana MCP allows OpsWorker's AI to query Grafana dashboards and metrics during investigations. They serve different purposes and can be used independently or together. See the Grafana integration page for setup instructions for each.
What is AI Memory?
AI Memory is OpsWorker's persistent context system. It captures facts about your environment (e.g., "we deploy on Fridays", "the payments database is owned by team X") and preferences for how the AI should behave. Memory is layered across personal (just you), cluster (shared with your team for a specific cluster), and organization (shared across all clusters) scopes. The AI uses memory automatically during investigations and chat to produce answers grounded in your operational reality. See AI Memory.