GoAlert

When hundreds of microservices trigger alerting storms during an infrastructure brownout, an on-call system cannot stall behind external rate limits. GoAlert processes notification cycles locally through a Go daemon backed by PostgreSQL, dispatching pages directly to on-call engineers without SaaS relay latency. Built by Target to handle enterprise retail operations, the platform executes escalation rules and rotation handoffs on private infrastructure under full local control.

Why GoAlert instead of Opsgenie

Opsgenie operates on-call schedules as a multi-tenant cloud service within Atlassian, routing alerts through remote infrastructure. That model suits teams wanting managed telecommunications, but it introduces external network hops, ingestion quotas, and a critical risk: when upstream cloud services degrade, the alerting system needed to coordinate recovery may stall. GoAlert eliminates external hops by running inside private clusters or on Linux nodes. Written in Go, its engine uses PostgreSQL transactions and advisory locks to evaluate incoming Prometheus or Grafana alerts in milliseconds, executing escalation policies without queue delays.

The operational tradeoffs reflect this divide. Opsgenie provides turnkey carrier routing and post-mortems linked to Jira. GoAlert focuses strictly on core paging mechanics—schedules, rotations, escalations, and notifications—without per-seat charges. While services like Opsgenie and PagerDuty impose monthly seat licenses that force teams to limit pager access, GoAlert lets organizations enroll every developer without incremental fees. Teams trade SaaS convenience for operational ownership: they maintain PostgreSQL availability and configure their own Twilio or SMTP credentials.

What it does well

• Compact single-binary deployment alongside PostgreSQL, avoiding distributed dependencies in critical alerting paths.
• Direct webhook ingestion from Prometheus Alertmanager, Grafana, Sensu, and custom HTTP clients without intermediate proxies.
• Multi-tier escalation policies with configurable delay thresholds that automatically progress from primary responders to secondary backups.
• Native heartbeat monitoring endpoints that trigger alerts whenever background cron tasks or worker daemons fail to check in.
• Bidirectional notifications across Twilio voice calls, SMS, and Slack, letting responders acknowledge or close alerts from phone prompts.