Pulsegrid / Integrations
Integrations
Forty-one integrations, all included on every plan. Each one is either a metrics scraper, an event source, or an alert destination.
Infrastructure
| Integration | Collects | Setup |
|---|---|---|
| AWS | CloudWatch metrics, ALB and RDS, ECS task events | Cross-account IAM role |
| Google Cloud | Cloud Monitoring metrics, GKE, Cloud SQL | Service account key |
| Azure | Monitor metrics, AKS, Azure SQL | App registration |
| Kubernetes | Node, pod and control plane metrics, events | Agent DaemonSet |
| Hetzner and OVH | Host metrics via the agent | Agent only |
Data stores and queues
| Integration | Collects |
|---|---|
| PostgreSQL | Query statistics, locks, replication lag, slow query log |
| MySQL and MariaDB | Performance schema, InnoDB metrics, slow query log |
| Redis | Keyspace, evictions, latency percentiles, replication |
| Kafka | Consumer lag per group, broker metrics, ISR state |
| RabbitMQ | Queue depth, unacked messages, connection churn |
| Elasticsearch | Cluster health, shard allocation, JVM heap |
| ClickHouse | Merges, parts, query log |
Alerting and workflow
| Integration | Direction | Notes |
|---|---|---|
| Slack | Outbound | Per-channel routing, threaded updates, resolve in place |
| PagerDuty | Outbound | Events API v2, auto-resolve when the alert clears |
| Opsgenie | Outbound | Priority mapped from alert severity |
| Microsoft Teams | Outbound | Adaptive cards |
| Generic webhook | Outbound | Signed with HMAC-SHA256, retried 5 times |
| GitHub Actions | Inbound | Deploy markers from workflow runs |
| GitLab CI | Inbound | Deploy markers from pipelines |
| LaunchDarkly | Inbound | Flag changes on the timeline |
| Terraform | Both | Manage alerts and dashboards as code |
No. All forty-one are on every plan including the free tier. Gating integrations behind a higher tier is a way of charging you twice for data you already sent.
Yes, and we build the ones several customers ask for. In the meantime, anything exposing a Prometheus endpoint can be scraped by the agent with four lines of config, which covers most of what people ask for.
Point the agent at a Prometheus endpoint, or push OTLP to the local agent on port 4317. Both are documented and neither needs anything from us.