The vendor-neutral autonomous incident-response layer for streaming video.
Streamwake investigates streaming failures, governs the remediation, and verifies that viewers actually recovered before the incident closes — end to end, on the same bus, for every incident the platform runs. Autonomous when safe, human-gated when necessary, auditable always. Vendor-neutral across encoder, CDN, DRM, and player — works across Mux, Bitmovin, AWS, Broadpeak, Wowza, any CDN, and PagerDuty, with no consolidation onto a Streamwake-owned monitoring suite.
Claim status labels: see definitions.
Seven stages, an explicit approval step before any write.
Read top to bottom on mobile, left to right on desktop. The agent proposes between Decide and Approve, but nothing writes to vendor config until a permitted operator signs off. The loop closes only when Verify confirms recovery.
Claim status labels: see definitions.
Detect. Investigate. Decide. Approve. Act. Verify. Learn.
The approved lifecycle is Detect → Investigate → Decide → Approve → Act → Verify → Learn. In a named scenario, each stage hands off to the next, with an explicit human Approve between Decide and Act. The evidence conditions determine what can be verified and learned.
Claim status labels: see definitions.
Detect
- A named evaluation scenario can collect player-QoE, encoder health, CDN egress, or DRM handshake evidence.
- Evidence is given a priority and stage within the configured review path, not treated as universal coverage.
- The documented surface can be compared with the metrics, logs, or traces available in that scenario.
Investigate
- A scoped review can join available encoder, CDN, DRM, and player evidence into a typed incident.
- Correlation can surface an affected surface, geography, and timeframe when those fields are evidenced.
- The documented workflow describes a structured view rather than a universal alert-ingestion promise.
Decide
- The documented evaluation path can classify a failure mode — such as manifest drift, edge brown-out, or DRM cold-start — with explicit confidence.
- It can propose a smallest-safe remediation candidate as a typed action within the reviewed scenario.
- A typed next action remains subject to the Approve step and does not imply direct access to vendor config.
Approve
- Approval is scoped to the reviewed scenario: permissions, channels, actions, regions, and blast-radius limits are explicit.
- Any evaluated write path requires the documented pre-authorization conditions before it can be considered for execution.
- Higher-risk writes remain paused for explicit operator approval before any write — governed, never ungoverned.
Act
- Describes the smallest safe remediation candidate for the reviewed scenario — for example, reroute egress, roll a flag, or re-package a title.
- Any evaluated execution is limited to explicitly whitelisted actions and a role-scoped blast radius.
- The documented path records a typed event scoped to the channel when that evidence surface is configured.
Verify
- The review can read the configured evidence path after an action and check whether the failure class changed.
- Evidence outside the configured window remains a question for the named scenario rather than a universal reopen rule.
- The lifecycle closes with a typed event only where the documented downstream surface is configured.
Learn
- A structured postmortem destination such as Slack, Linear, or PagerDuty is an evaluation-scoped output, not a universal integration claim.
- The reviewed record can capture what happened, what was tried, and what changed when those fields are available.
- Replay artifacts are a scenario-specific evidence question, not a general delivery commitment.
What the agent is allowed to touch.
By default Streamwake reads signals rather than writes to vendor config. Writes happen only against actions you have whitelisted, are scoped to the channel, are gated behind the explicit Approve stage between Decide and Act, and are recorded as typed, auditable events you can replay.
Claim status labels: see definitions.
- A configured Datadog event, New Relic event, or Grafana annotation as an evaluated output.
- Rerouting egress on an explicitly whitelisted Metro or region in the named scenario.
- Re-packaging or re-keying a title through an authorized packager credential, scoped to the channel.
Bring us your last streaming incident. A walkthrough, on us.
You share the timeline — the affected geography, the symptom, and the moment it cleared. A human-led review can assess the available evidence across encode / edge / DRM and discuss a typed remediation candidate without promising a root-cause finding, production action, or customer outcome.