How AI agents monitor cloud provider release notes
An agent monitors cloud release notes by calling a wire that reads AWS what's-new, Azure service updates, GitLab and Chrome release posts at runtime, filtered with src. Release notes announce features, deprecations and retirements. They are a separate stream from status pages, which report incidents.
/mcp. Nothing here needs an API key or a signup.
Those two streams get merged constantly, and merging them is how an agent ends up reporting a feature launch as an outage.
A launch, a retirement and an incident are three things
A release note says something changed on purpose. An incident report says something broke. The tell is tense and intent: release notes are written in advance and scheduled, incident reports are written under pressure and updated until resolved.
Inside release notes there is a second split that matters more to an agent than the feature announcements do. A retirement or deprecation notice carries a date in the future on which existing, working code stops working. That is the only category in this stream with a deadline attached.
| Item type | What it means | Urgency |
|---|---|---|
| Generally available | A feature is now usable | Optional |
| Public preview | A feature exists, terms differ | Optional |
| Retirement | Something you use stops on a date | Deadline |
| Incident | Something is broken now | Immediate |
The first three come from this wire. The fourth is the outage wire, covered in how agents check whether AWS or OpenAI is down.
Azure makes the split easy to query because the status is written into the item title. A live read returned titles prefixed [Launched], [In preview] and Retirement:, including one reading "Retirement: Nested confidential (cc_v5) VMs will be retired on September 1, 2026". A q=retirement query against this wire works because the publisher put the word there, not because we inferred it.
Volume is the design constraint
A live read of the AWS what's-new feed returned 100 items, of which 47 were published inside 3 days. That single publisher generates more items per week than most government wires generate per quarter.
Retention here is 3 days with a 3 day default lookback. That is short on purpose. A longer window on this wire would return more than an agent can read in one call and would still miss AWS items that had already fallen off the publisher's own feed. The route is for noticing what changed this week. Each provider maintains a permanent archive of its own for anything older.
The four publishers behind the wire are AWS, Azure, GitLab and Chrome, filtered with src=aws, src=azure, src=gitlab and src=chrome. Chrome is the predictable one: Google states that Chrome Stable gets a major release every four weeks and minor updates every two to three weeks, rolled out in stages starting at 1 to 5 percent of users. An agent can plan around that cadence. AWS and Azure post continuously and cannot be planned around, only queried.
The call an agent makes
# discovery only, no items: browse every wire, its filters and its price claude mcp add --transport http botwire https://thebotwire.com/mcp # paid, $0.005: deprecations and retirements across all four publishers curl -i "https://thebotwire.com/cloud/releases?q=retirement&since=3d&limit=25" # one provider, one service area curl -i "https://thebotwire.com/cloud/releases?src=aws&q=lambda&since=3d"
Parameters are src (aws, azure, gitlab, chrome), q, since and limit (1 to 50). The call costs $0.005 in USDC on Base, with no API key, no signup and no subscription. Search is ranked keyword matching, so service names outperform a phrased question. Every route, filter and price is machine-readable at /llms-full.txt.
Managed platform releases and open-source project releases are different questions with different publishers. Version tags for runtimes and frameworks are covered in how agents monitor open-source releases of their dependencies.
FAQ
Does this wire cover Google Cloud release notes?
No. The four publishers are AWS, Azure, GitLab and Chrome. Google Cloud product release notes are not in this wire, though Google Cloud incidents do appear in the separate outage wire. An agent answering a multi-cloud question should name which providers it actually searched rather than implying full coverage.
Can an agent get the exact retirement date from a wire item?
Usually yes, because providers put the date in the title or the summary, as in the Azure example above. It is still the announcement rather than the authoritative deprecation schedule. When a date is load-bearing for a migration plan, the honest pattern is to quote the announcement, link it, and fetch the provider's own deprecation page before acting on it.
Why is the window only 3 days?
Because AWS alone published 47 items in 3 days on a live read, and a wider window turns a single call into a wall of text. The wire exists to notice that something landed. Each provider's archive exists to retrieve something you already know landed.
How fresh is the cloud release wire?
It sits in the sparse ingest tier and is polled at most every 15 minutes. Given that these publishers post throughout the working day, the practical lag between a provider posting and the item being queryable is minutes rather than hours. The observed ingest interval and last poll timestamp are published live at /health, which is the ground truth for any cadence claim.
Sources
Chrome's release cadence of a major release every four weeks, minor updates every two to three weeks and staged rollout beginning at 1 to 5 percent of users is documented by Google in Chrome release channels. The AWS item counts above come from the provider's own recent announcements feed, read on 7 August 2026. Poll cadence and per-call pricing for this service are published live at thebotwire.com/health.
Related: How do AI agents check whether AWS or OpenAI is down? · How do AI agents monitor open-source releases of their dependencies?