Here’s a roundup of the new features and improvements we shipped over the past month.

Checkly MCP Server

We released our first version of the Checkly MCP! You can read the announcement or install it straight away using its URL:

https://api.checklyhq.com/mcp

You’ll find more specific instructions for each supported client, as well as limitations in our documentation.

Checkly CLI v8.10.0: What’s New

Upgrade to v8.10.0 to access the latest CLI features and improvements released over the past month.

Deploy larger CLI projects reliably (CLI v8.10.0)

The CLI now supports large deployments more reliably, including projects with thousands of resources. For longer deployments, the CLI displays progress based on the resources being created or updated. It also handles multiple deployments started at the same time:

• By default, a new deployment waits for the current deployment to finish.

• Use -cancel-in-progress-deployment to cancel the running deployment and start the new one immediately.

Result assets, straight from the terminal (CLI v8.8.0)

No more digging through nested check-result JSON. List and download the artifacts of any check run or test-session result - logs, Playwright traces, videos, screenshots, pcaps, reports:

• checkly assets list --result-id <id> --check-id <id> (or --test-session-id) — filter by --type/--asset, view as table/tree/json

• checkly assets download <same as above> --type all — pulls everything into ./checkly-assets/…, with --force / --skip-existing

→ Learn more in our docs.

Drill into a single test-session result (CLI v8.7.0)

Jump straight to one result's logs and timing inside a session with checkly test-sessions get <id> --result <result-id>. Pairs with checkly test-sessions list to find the session first.

→ Learn more in our docs.

Delete a check by ID (CLI v8.8.0)

Delete a check by ID with checkly checks delete <id>. Confirmation prompt by default (or --force), plus --dry-run to preview. Note that checks managed by a CLI project get recreated on the next deploy, remove those from your project code instead.

→ Learn more in our docs.

Manage account members (CLI 8.7.0)

Use checkly account members update / delete to manage account members. Beyond listing, you can now change a member's role (--role) or remove them, by email or user ID. Confirmation prompt and --dry-run included.

→ Learn more in our docs.

Tag Manager: Manage tags in one place

The Tag Manager is now available in the top level menu under Configuration, making it easier to manage tags across your entire account.

You can now:

• Rename existing tags: Updates the tag everywhere it is used, including checks, groups, and maintenance windows.

• Delete unused or incorrect tags: Clean up typos and outdated tags without updating each resource individually.

• Add tag descriptions: Give Rocky AI additional context it can use when generating root cause analyses.

The Tag Manager is currently available in the Checkly UI only. Head to the Checkly Webapp to try it out now.

Developer Experience

Checkly Slack app

You can now install the Checkly Slack app to receive alerts and rerun checks or AI analyses directly from Slack. The app replaces webhook-based Slack alerts and is available through a new CLI construct, with Terraform support coming soon. Learn more on the in-app installation page or in the docs.

Cancel Agentic Check runs

You can now cancel Test prompt runs, Test session runs, and Check runs for Agentic Checks from both the API and UI.

Rocky AI RCA now uses OTel traces and the last passing result

To improve root cause analysis, Rocky AI now automatically looks for related OTel traces and compares the failure against the most recent passing result. This gives Rocky more context to identify regressions and explain what changed. Read the full changelog for more information.

Trace IDs are now included in check result API responses

When using Checkly OpenTelemetry Traces, the Check Run Result now returns the corresponding Trace ID through the API:GET /v1/check-results/{checkId}/{checkResultId}. You can use this ID to correlate a check run with its trace in Checkly or your observability backend.

HTTPS record support for DNS monitors

DNS monitors now support the HTTPS resource record type. Learn more in the DNS monitor docs.

Happy monitoring!

Questions or feedback? Join our Slack community.

AI agents can now connect directly to Checkly through the Model Context Protocol (MCP).

The new Checkly MCP Server gives your agentic client of choice a secure, OAuth-based access to your Checkly account, so your agent can inspect live monitoring data, understand check health, investigate recent failures, and help with incident workflows without leaving your conversation.

With the Checkly MCP Server, your agent can:

• List accounts and understand available Checkly features

• Inspect check status, recent results, test sessions, and result assets

• Read and trigger Rocky AI root cause analyses

• Trigger existing deployed checks on demand

• Read status pages and manage status page incidents

• Manage account-level environment variables, with secret values protected

This is designed for live Checkly account operations. For creating, editing, testing, or deploying Monitoring as Code projects, keep using Checkly Skills and the Checkly CLI. The MCP server runs remotely, so it cannot access your local project files.

You can connect Codex, Claude Code, Cursor, VS Code, or any MCP client that supports Streamable HTTP and OAuth.

Learn how to connect your client in the MCP Server docs:
https://checklyhq.com/docs/ai/mcp-server

Happy monitoring!

Questions or feedback? Join our Slack community.

We just added two new sources of data Rocky AI can pull from to create a root cause analysis for your failing checks, monitors and test results.

Open Telemetry Traces

If you are using Checkly Traces, Rocky AI now automatically finds, queries and evaluates relevant traces linked to your failures. The example below is from Checkly’s own backend infrastructure.

• Our API check that checks if our customer facing Prometheus endpoint works, failed with a 500 error.

• The OTeL trace indicates this was actually do to our Clickhouse server returning a 500 errors, immediately telling our on-call team where to start looking.

• Without the trace, we would have to look at all the various logging, error tracking and other tools that are integrated into the various middleware and infrastructure this request passes through (our load balancer, REST API server, Redis datastore etc.)

This feature is now live for all Checkly Traces users. Checkly Traces and Rocky AI Root Cause Analyses are part of the Checkly Resolve package. 10 Rocky AI root cause analyses are part of the free Checkly Hobby plan.

Last passing result

Rocky AI now automatically searches for and interprets a last passing result for your failing check, monitor or check result. This is useful to more clearly indicate what the nature of a regression is: specifically for more “verbose” checks like Playwright, Browser and Multistep checks.

It can indicate for instance that a network error is transient, not persistent, as it can see that an earlier call — 2 minutes earlier in the example below — passed without failure.

Two caveats:

1. A last passing result for monitors and checks might not always exist: the check might start failing immediately after creation.

2. A last passing result for test session might not exist because it is the first run, or when the test name was changed, the individual spec was removed from a test suite etc.

This feature is now live for all Rocky AI Root Cause Analyses users.

Happy monitoring!

Questions or feedback? Join our Slack community.