Comparison — vendor pages read 2026-11-11
Observability Unified vs Sentry
Sentry covers errors, tracing, logs, session replay on web and mobile, and Agent Tracing, as a SaaS or as a Fair Source self-hosted install.
Observability Unified is a self-hosted platform that connects traces, logs, replay, profiles, and AI agent activity through shared IDs, and gives AI agents evidence with citations through MCP. Each row below says Yes, Partial, or No for both products and links to the page that supports it.
Where Observability Unified differs
- MCP write tools off unless enabled. Writes register only with OBS_MCP_ENABLE_WRITES=true
- Compacted evidence for AI agents. CCR bundles collapse repeated logs to exemplars within a token budget; refs expand them
- Records what an agent's tool call changed. Before, after, and diff values your app attaches to a tool call, stored as sent
- Production traces become eval cases. Save as eval case from a trace or run; candidate runs compared step by step (HTTP API)
- Shared-dependency contention vs code regression. Shared-degradation verdicts separate contention from the new version
- Root cause states its level of proof. Proof tier from unverified locus to counterfactual verified, kept apart from link certainty
- Fix verified by re-running the failing case. Same eval input fails on the baseline and passes on the candidate
- Remediation behind approval or a policy gate. Approval queue or single-use grant; your runner executes; a watchdog checks recovery
- Matches new incidents to past ones. Prior findings matched by fingerprint appear in incident evidence bundles
- Distributed tracing (OTLP). OTLP over HTTP; front gRPC with a Collector
- Run the full platform yourself. Docker image, Node + Postgres/S3, or Cloudflare
Where Sentry offers more
- Write your own queries over telemetry. Trace Explorer query builder with aggregates and group-by on spans
- Vendor-hosted cloud service. Sentry SaaS
- Built-in single sign-on. SAML2 on Business and Enterprise
Capability by capability
Investigating with AI agents
| Capability | Observability Unified | Sentry |
|---|---|---|
| First-party MCP server | Yes Local stdio server, @obsunified/mcp-server our docs | Yes Hosted with OAuth; local stdio through the CLI mcp.sentry.dev |
| MCP write tools off unless enabled | Yes Writes register only with OBS_MCP_ENABLE_WRITES=true our docs | Partial Write skills are chosen at OAuth consent cli.sentry.dev |
| MCP results link to source records | Yes Evidence bundles with citations; dashboard deep links when OBS_DASHBOARD_URL is set our docs | Yes Tool responses include Sentry URLs github.com |
| Agent tool calls alongside app traces | Yes Agent Action Graph links tool calls to traces our docs | Yes Agent runs shown with the app's trace docs.sentry.io |
| Compacted evidence for AI agents | Yes CCR bundles collapse repeated logs to exemplars within a token budget; refs expand them our docs | Partial MCP trace tool returns a condensed overview; search_traces fetches the rest github.com |
| Agent skills for coding agents | Yes Instrument and investigate skills, released as .skill bundles our docs | Yes Official Sentry plugins for Claude Code, Cursor, Codex ship agent skills github.com |
Observing your own AI agents
| Capability | Observability Unified | Sentry |
|---|---|---|
| Records what an agent's tool call changed | Yes Before, after, and diff values your app attaches to a tool call, stored as sent our docs | Partial Agent tool calls recorded with inputs and outputs in traces docs.sentry.io |
| MCP calls recorded with trace context | Yes MCP spans join the action graph; an optional allow-listed envelope records the context our docs | Yes MCP server monitoring traces tool, resource, and prompt calls as spans docs.sentry.io |
| Production traces become eval cases | Yes Save as eval case from a trace or run; candidate runs compared step by step (HTTP API) our docs | Partial Shows LLM-as-judge decision spans in traces docs.sentry.io |
Change-aware investigation and remediation
| Capability | Observability Unified | Sentry |
|---|---|---|
| Compare releases and flag variants | Yes Error-rate lift per version and flag variant, normalized for traffic share our docs | Yes Release metrics compared with all releases docs.sentry.io |
| Shared-dependency contention vs code regression | Yes Shared-degradation verdicts separate contention from the new version our docs | Partial Seer can point to a backend service as the cause docs.sentry.io |
| Root cause states its level of proof | Yes Proof tier from unverified locus to counterfactual verified, kept apart from link certainty our docs | Partial Seer names the most likely root cause docs.sentry.io |
| Fix verified by re-running the failing case | Yes Same eval input fails on the baseline and passes on the candidate our docs | Partial Seer PRs iterate on failing CI checks docs.sentry.io |
| Remediation behind approval or a policy gate | Yes Approval queue or single-use grant; your runner executes; a watchdog checks recovery our docs | Partial Autofix opens PRs; automation stop point is configurable docs.sentry.io |
| Matches new incidents to past ones | Yes Prior findings matched by fingerprint appear in incident evidence bundles our docs | Partial Resolved issues auto-flag as Regressed; AI grouping merges similar errors docs.sentry.io |
Signals
| Capability | Observability Unified | Sentry |
|---|---|---|
| Distributed tracing (OTLP) | Yes OTLP over HTTP; front gRPC with a Collector our docs | Partial Tracing; OTLP traces in open beta docs.sentry.io |
| Structured logs | Yes Trace-correlated our docs | Yes Trace-connected docs.sentry.io |
| AI / LLM observability | Yes AI calls, agent runs, evals our docs | Yes Agent Tracing and the Seer debugging agent docs.sentry.io |
| Session replay | Yes rrweb, browser DOM only our docs | Yes Web and mobile docs.sentry.io |
| Browser click linked to its backend trace | Yes x-obs-interaction links a click to its trace, logs, and replay (browser SDK or header) our docs | Yes Click spans parent their requests (experimental) docs.sentry.io |
| CPU profile from a trace | Yes When profiling is enabled and samples carry the trace ID our docs | Yes Span details show the profile docs.sentry.io |
| One ID chain from user to agent action | Yes user_id → session_id → interaction_id → trace_id → span_id → action_id our docs | Yes Replays link to traces; agent conversations link to traces and users docs.sentry.io |
| Write your own queries over telemetry | Partial Search and filters; no query language or SQL editor our docs | Yes Trace Explorer query builder with aggregates and group-by on spans docs.sentry.io |
| Points out gaps in your instrumentation | Yes Flags spans with unexplained self-time in the waterfall and in get_trace our docs | Yes Trace view marks orphan traces and no-instrumentation spans docs.sentry.io |
Deployment and access
| Capability | Observability Unified | Sentry |
|---|---|---|
| Run the full platform yourself | Yes Docker image, Node + Postgres/S3, or Cloudflare our docs | Partial Self-hosted for low-volume deployments github.com |
| Vendor-hosted cloud service | No Self-hosted only our docs | Yes Sentry SaaS sentry.io |
| Built-in single sign-on | No Front it with Cloudflare Access or oauth2-proxy our docs | Yes SAML2 on Business and Enterprise docs.sentry.io |
Choosing between them
Choose Sentry if you want errors, session replay including mobile, and traces without running infrastructure.
Choose Observability Unified if you want AI agents and engineers to investigate the same production evidence, with the telemetry kept on infrastructure you run. It is early-stage software suited to single-team deployments; the documented trade-offs include no built-in SSO, OTLP over HTTP only, and browser-only session replay.
How these cells were checked
Yes and Partial cells cite the vendor's own docs, pricing, product page, or official repository. No is used only where the cited page says so. Not marketed† means we read the vendor's product pages and found no such capability; it is an inference, not a vendor statement. Not stated means the vendor's docs do not say either way.
Rows for MCP write defaults, MCP links, agent tool calls, CPU profiles, self-hosting, and SSO were read on 2026-11-10 (ledger with excerpts). Rows for release and flag comparison, proof of cause, fix verification, gated remediation, click to trace, and dependency contention were read on 2026-11-11 (ledger with excerpts). Rows for compacted evidence for agents, agent skills, one ID chain, and instrumentation gaps were read on 2026-11-11 (ledger with excerpts). Rows for tool-call side effects, MCP call recording, production traces to eval cases, incident memory, and ad-hoc queries were read on 2026-11-11 (ledger with excerpts). The MCP server row was read on 2026-11-10 (ledger), and the remaining rows on 2026-10-13 (ledger). Vendors change their products often; if a cell is out of date, send a correction. The full comparison covers all nine vendors with excerpts.