[codex] docs: explain k8s proxy benefits#847
[codex] docs: explain k8s proxy benefits#847officialasishkumar wants to merge 9 commits intomainfrom
Conversation
Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
284d697 to
9c4c942
Compare
Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
There was a problem hiding this comment.
Thank you for this deep-dive on the Kubernetes Proxy benefits! The updated section is significantly more useful than the bullet-list version - walking through each capability with concrete detail (auto-replay cadence, noise classification buckets, MCP tool names, CI workflow steps) gives teams exactly the context they need to decide whether to adopt the proxy.
Issues to be Considered:
-
The heading change removes backtick formatting from
`keploy enterprise`(now rendered as plain textkeploy enterprise directly). Since this is a CLI command name, keeping it in code formatting helps readers scan the page. Worth restoring. -
Section 5 (Smart test set) is labelled "Upcoming" in the heading but the body text switches between present and future tense in a way that may confuse readers about what is actually available today vs. on the roadmap. Consider adding a short callout box at the top of that section like
> **Status: Not yet active in the current record/start path." so it's unambiguous. -
This PR modifies
k8s-proxy-api.md, which is also the target of PR #845 (the revert PR currently open). These two will produce a merge conflict. Please coordinate with the maintainers on which lands first, or rebase onto the result of #845 before this is merged. -
The new
k8s-arch.pngimage is added but not tested in both Docusaurus light and dark themes. The image alt text (How Keploy works on Kubernetes architecture overview) is descriptive which is great - just make sure the image itself renders legibly in both color modes. -
The Local CI replay section references
keploy test sandboxas the command but the earlier docs usekeploy test. Ifsandboxis a separate subcommand, please add a note pointing readers to the CLI reference so they know where to find the flags.
Overall this is a substantive and well-written addition. Sorting out the merge conflict with the revert PR and clarifying the upcoming feature boundary are the main things to address before this is ready.
| Each test case is exercised once against a freshly rolled Pod and classified into one of three buckets: | ||
|
|
||
| - **Pass:** kept as a real, stable test case. | ||
| - **High-risk failure:** marked as failed (a real regression to investigate). |
There was a problem hiding this comment.
do we do this? isnt it just pass or fail
There was a problem hiding this comment.
yes, we do have this.. we added this in report as well that gets generated after running keploy test
charankamarapu
left a comment
There was a problem hiding this comment.
Please address the comment rest looks good
| > [!NOTE] | ||
| > This documentation covers **local Kind cluster** setup. Documentation for **hosted cluster** setup is coming soon. | ||
| ## How Keploy works on Kubernetes |
There was a problem hiding this comment.
the arch picture is missing few things.
- sidecar should be inside pod
- Only self hosted arch is present
- Secrets missing
- MITM proxy hope reader doesn't get confused with other product - https://github.com/mitmproxy/mitmproxy
- why does Kubernetes api-server has test case, mocks mentioned.
There was a problem hiding this comment.
updated the image
| - The **Kubernetes Proxy REST API** ([Endpoint reference](#endpoint-reference)) handles in-cluster operations such as starting/stopping recording, kicking off replay, fetching session status, and reading logs or reports. All routes outside `/healthz` and the admission webhook sit behind shared-token Bearer auth. | ||
| - The **Keploy API server MCP endpoint** exposes higher-level tools for AI coding tools, including Claude Code, Cursor, Windsurf, and VS Code. This is how an AI agent in your editor authors test suites, runs replays, and scaffolds CI pipelines without you copy-pasting curl commands. | ||
|
|
||
| The MCP surface includes around a dozen tools. The headline ones: |
There was a problem hiding this comment.
when it comes to just k8sproxy these and integration tests these are not useful. Mentioning them will confuse user.
Refine the documentation for the Kubernetes Proxy REST API, emphasizing that all console actions are accessible via REST calls. Removed details about the MCP server and its tools, focusing instead on the available endpoints for recording, testing, and status monitoring. Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
b9c0e21 to
dbee596
Compare
There was a problem hiding this comment.
Pull request overview
Expands the Kubernetes Proxy REST API documentation to explain why teams should prefer the in-cluster proxy model over running keploy enterprise directly in each application Pod, and adds a Kubernetes architecture overview image to the Kubernetes setup guide.
Changes:
- Added a detailed “Why Kubernetes Proxy” benefits section (auto-replay, deduplication, automation surfaces, schema management, upcoming smart test set, and local CI replay).
- Inserted a Kubernetes architecture overview image into the Kubernetes setup guide.
Reviewed changes
Copilot reviewed 2 out of 3 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| versioned_docs/version-4.0.0/running-keploy/k8s-proxy-api.md | Replaces the brief benefits list with a structured, detailed rationale for using the Kubernetes Proxy, including operational/behavioral edge cases. |
| versioned_docs/version-4.0.0/keploy-cloud/kubernetes.md | Adds a “How Keploy works on Kubernetes” section with an architecture overview image. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Replaced the existing k8s-arch.png with a new version to enhance visual representation of the architecture. Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
aef783a to
da812aa
Compare
Updated the capitalization of "Keploy Enterprise" in the Kubernetes Proxy API documentation for consistency and clarity. Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
d38398b to
b8725ca
Compare
Drop the Local CI replay point from the v4.0.0 k8s proxy API docs to keep the page aligned with the requested scope. Made-with: Cursor Signed-off-by: Asish Kumar <officialasishkumar@gmail.com>
b8725ca to
0bb8a91
Compare
What has changed?
This PR expands the Kubernetes Proxy API docs to explain why teams should use the in-cluster proxy instead of wiring
keploy enterprisedirectly into each application Pod.The update documents the main user-facing benefits: auto-replay during recording, release-aware replay context, capture-time deduplication, REST and MCP automation surfaces, schema generation and schema-conflict handling, the upcoming smart test set model, and local CI replay. It also clarifies edge cases from the implementation, including low-risk noise handling, re-record suite-linking gates, schema insert-vs-update behavior, and the current status of historical/smart-test-set consolidation.
This PR also adds a Kubernetes architecture overview image to the Kubernetes setup guide.
This PR does not resolve a linked issue.
Type of change
How Has This Been Tested?
git diff --check -- versioned_docs/version-4.0.0/running-keploy/k8s-proxy-api.md.git diff --check -- versioned_docs/version-4.0.0/keploy-cloud/kubernetes.md static/img/k8s-arch.png.vale versioned_docs/version-4.0.0/keploy-cloud/kubernetes.md; it only reports a pre-existing frontmatter issue outside this PR's changed lines.npm run buildsuccessfully after adding the image; Docusaurus generated static files inbuild.npm run serve -- --host 127.0.0.1 --port 3010successfully earlier on this branch; Docusaurus served the build athttp://127.0.0.1:3010/docs/, then the local server was stopped.The visual asset added by this PR is
static/img/k8s-arch.png.Checklist: