Why developers leave Sentry: pricing on high-error-volume apps adds up, logs (vs errors) are a newer product less mature, and dashboards less customizable than Datadog/Grafana. Teams whose Sentry use grew toward APM-style needs evaluate broader observability tools.
Datadog covers the breadth Sentry doesn't — APM, logs, infra, RUM, security. The pane-of-glass story is real for teams that need infrastructure visibility alongside code-level errors. The trade-off is opaque per-feature pricing.
Grafana Cloud free tier + Grafana for self-host alternative. OpenTelemetry-native. OSS-first. For cost-conscious teams or those needing self-host Grafana's LGTM stack covers the broader observability that Sentry alone doesn't.
Honeycomb wins for distributed-systems debugging — high-cardinality queries Sentry can't answer directly. BubbleUp anomaly attribution. OpenTelemetry-native. Right alternative if your Sentry use was creeping toward "why is this slow" territory.
Most reviewers supplement. Keep Sentry for code-level errors and release health (it remains best-in-class). Add Grafana for infra metrics, Honeycomb for distributed-systems debugging. Migrating off Sentry entirely loses functionality the alternatives don't replicate.
Error grouping defaults are better. Release tracking is first-class. Session Replay for understanding bug context. Datadog has all of these as features but Sentry's defaults route closer to engineering workflows. For code-first teams Sentry stays the right tool.
For new instrumentation: yes. Vendor-portable. Sentry supports OTel inputs. Most modern observability tools either build on OTel (Honeycomb, Grafana Tempo) or accept OTel as primary input. The instrumentation-tax of vendor-specific SDKs has diminishing returns.
