Data Sources

Connect your observability, infra, and CI/CD tools to give the SRE AI Agent access to your stack.

Data sources are the tools and services the SRE AI Agent connects to during incident investigation. Each data source provides a different type of signal — logs, metrics, traces, deployments, code changes, or infrastructure state.

How to Connect a Data Source

1. Go to SRE Agent Integrations in the Pagerly dashboard

2. Click Add Integration

3. Select the data source you want to connect

4. Enter the required credentials (API key, token, URL, etc.)

5. Select the team this data source belongs to

6. Add an optional context prompt to help the agent understand your setup

7. Click Save

All credentials are encrypted at rest and only used by the SRE Agent during active investigations. We recommend creating read-only API keys with minimal permissions.

How Mapping Works

Each data source is mapped to a specific on-call team. When someone triggers @Pagerly assist in a channel, the agent looks up which team owns that channel and only queries the data sources mapped to that team.

This means:

• Team A (Backend) might have Datadog, Kubernetes, PostgreSQL, and GitHub connected

• Team B (Frontend) might have Sentry, Vercel, and GitHub connected

• Team C (Platform) might have AWS, CloudWatch, Grafana, and ArgoCD connected

When an incident is reported in Team A's channel, the agent only queries Datadog, Kubernetes, PostgreSQL, and GitHub — not Vercel or Sentry.

Available Data Sources

Observability & Monitoring

Query logs, metrics, traces, and errors from your monitoring stack.

Data Source	What the Agent Queries
Grafana	Dashboards, annotations, Prometheus metrics
Coralogix	Logs, metrics, traces
Datadog	Logs, metrics, APM traces, monitors
Prometheus	Metrics, firing alerts
Grafana Tempo	Distributed traces (TraceQL)
Loki	Logs (LogQL)
VictoriaLogs	Logs (LogsQL)
Elasticsearch	Logs, search data
OpenSearch	Logs, search data
Splunk	Logs, metrics (SPL)
Jaeger	Distributed traces
New Relic	APM, logs, infrastructure (NRQL)
Honeycomb	High-cardinality data, traces, SLOs
Sentry	Application errors, stack traces
CloudWatch	AWS logs, metrics

Infrastructure

Check the health and state of your infrastructure.

Data Source	What the Agent Queries
Kubernetes (Direct)	Pods, deployments, events, logs
Kubernetes (Agent)	Same as above, via in-cluster agent
AWS	EC2, ECS, Lambda services
Google Cloud (GCP)	GCE instances, GKE clusters, Cloud Monitoring
Azure	Resources, AKS, Azure Monitor, Activity Log
Docker	Containers, health, restart counts, logs
Helm	Helm releases, revision history

Databases & Message Queues

Monitor database health and queue depths.

Data Source	What the Agent Queries
MongoDB Atlas	Cluster health, processes, alerts
PostgreSQL	Active queries, locks, replication lag
Kafka (Confluent)	Topics, consumer groups, consumer lag
RabbitMQ	Queue depths, connections, dead-letter queues

CI/CD & Deployments

Correlate incidents with recent deployments and builds.

Data Source	What the Agent Queries
ArgoCD	GitOps app sync status, deploy history
Jenkins	Build status, console logs
Vercel	Deployments, functions, logs

Code & SCM

Search code and recent changes that might have caused the incident.

Data Source	What the Agent Queries
GitHub	Code, PRs, commits, deployments
GitLab	Code, merge requests, pipelines

Incident Management & Ticketing

Pull context from your incident management tools.

Data Source	What the Agent Queries
PagerDuty	Incidents, alerts, escalations
OpsGenie	Alerts, on-call schedules
ServiceNow	Incidents, change requests, CMDB
Jira	Issues, tickets, post-mortems
Confluence	Runbooks, documentation
ClickUp	Tasks, incident tracking

Other

Data Source	What the Agent Queries
Web Search	Stack Overflow, GitHub Issues, known fixes
Amplitude	Product analytics, user activity

Can't find your tool? Contact us at [e-mail] — we add new integrations regularly.