Atlassian is shutting Opsgenie down entirely by April 2027, forcing every team still on it into an unplanned migration, while PagerDuty keeps billing $21-$49 per responder each month and caps its free tier at one escalation policy, one schedule, and 100 SMS/voice notifications before overage fees apply. Grafana OnCall, long the default open-source escape hatch from both, had its OSS repository archived in March 2026, cutting off the Cloud Connection relay it used for SMS and voice paging. Self-hosting swaps that per-seat math for infrastructure you already run: GoAlert ships as a single Apache-2.0 binary with direct Postgres access to its own schedule and escalation-policy tables, while Keep's YAML workflows execute arbitrary Python or webhook steps against your alert data with no per-workflow metering. OneUptime bundles on-call scheduling into the same Helm chart as its Postgres- and ClickHouse-backed monitoring stack, and Robusta's rule engine runs inside your own Kubernetes cluster to pull pod logs straight from the API server. Crucially, Prometheus Alertmanager provides the upstream routing, grouping, and silencing layer, but since it does not handle on-call schedules or SMS/voice paging on its own, teams pair it with tools like GoAlert or Keep for complete incident management.
| Name | Main Category | Sub-Category | Target Alternative | Free Tier Summary | Self-Hostable? | Short Description | Primary Link |
|---|---|---|---|---|---|---|---|
GoAlert | Product Analytics & Monitoring | On-Call Scheduling & Escalation | PagerDuty / Opsgenie | 100% Free - Self-Hosted (Apache 2.0) | Yes - Binary | A single Apache-2.0 Go binary built by Target that handles on-call scheduling, escalation policies, and Twilio-backed SMS/voice notifications against its own Postgres schema. Ships with a one-command demo Docker image so you can test full escalation chains before wiring up production Twilio credentials. | |
OneUptime | Product Analytics & Monitoring | Incident Response & Observability Platform | PagerDuty / Opsgenie / Statuspage | Free (Apache 2.0) - self-hosted Community edition; paid Enterprise tier adds hardened images and support | Yes - Docker / Kubernetes | An Apache-2.0 platform bundling uptime monitoring, status pages, on-call scheduling, and OpenTelemetry-based APM/logging into one self-hosted stack. Deploys via Helm for production, replacing PagerDuty, Statuspage, and Datadog with a single Postgres/ClickHouse-backed system you own. | |
Prometheus Alertmanager | Product Analytics & Monitoring | Alert Routing & Deduplication | Opsgenie / PagerDuty (Alert Routing Layer) | 100% Free - Self-Hosted (Apache 2.0), CNCF-graduated project | Yes - Binary | The CNCF-graduated routing and deduplication engine that sits between Prometheus (or any webhook source) and your paging tool, grouping related alerts and silencing known noise before they reach an on-call engineer. Ships as a static Go binary configured entirely through a version-controlled YAML routing tree. | |
Keep | Product Analytics & Monitoring | AIOps & Alert Correlation | Opsgenie / PagerDuty / Datadog | 100% Free - MIT Open-Source Core; AI Correlation, SSO/RBAC, & HA gated behind Enterprise tier | Yes - Docker | An MIT-licensed alert management platform that deduplicates and unifies alerts from 110+ monitoring tools (Datadog, Grafana, Prometheus), executing GitHub-Actions-style YAML workflows for auto-remediation and routing. Rule-based correlation and unlimited integrations are fully self-hostable, while LLM/AI-driven correlation requires an Enterprise license. | |
Robusta | Product Analytics & Monitoring | Kubernetes Alert Enrichment & Auto-Remediation | Opsgenie (Kubernetes Alert Enrichment) | Free (MIT) - self-hosted rule engine; optional paid SaaS UI for multi-cluster management | Yes - Kubernetes | An MIT-licensed rule engine that runs inside your Kubernetes cluster, enriching Prometheus alerts with pod logs and graphs before forwarding them to Slack, PagerDuty, or Opsgenie. Auto-remediation rules can restart CrashLoopBackOff pods or scale deployments without a human opening a laptop. |
Airtable Free allows 100 automation runs a month; Zapier Tables Free stores 2,500 records. Compare five self-hosted tables with native webhooks.
Astro and MWAA bill for uptime, not runs. Compare Airflow, Dagster, Prefect, Mage and DolphinScheduler as self-hosted pipeline orchestrators.
Zapier Agents meters every activity and StackAI caps free seats. Compare five self-hosted LLM chain builders with Docker and zero per-run billing.
Workato and Tray.io meter every workflow step as a task. Compare five self-hosted integration engines that run in your VPC with no per-step meter.
Step Functions bills every state transition and Retool Workflows meters monthly runs. Five self-hosted engines run code-first workflows with no run caps.
Pipedream bills credits by compute time and has no self-hosted edition. These five engines run code-first workflows, webhooks and cron on your own servers.
Softr meters app users and Zapier Interfaces meters pages. Four open-source builders run client portals and form apps on your own Docker host.
Retool now reserves self-hosting for Enterprise contracts. Here are five open-source builders you can run on your own Docker host with no per-seat license.
Zapier's free plan blocks webhooks and bills every action as a task. Compare six self-hosted engines that catch webhooks and run unlimited multi-step flows.
Loom's free plan stops at 25 videos and 5-minute recordings. Self-host Cap, SendRec, Clips or PeerTube and record without per-creator limits.
Miro's free tier allows three editable boards, then bills per member. Five self-hosted canvases for diagrams, sprint sketches, and live brainstorms.
Basecamp only runs as SaaS. These five open-source tools rebuild its message boards, to-dos, chat, and file storage on servers you control.