docs: clarify private link traffic imbalance for all node types including APM#5801
Conversation
…ding APM Expands the 'Before you begin' section to: - Explain that traffic imbalance affects all node types (ES coordinating nodes and APM/Integration servers), not just coordinating nodes - Describe the AWS NLB flow-hash routing behavior as the root cause - Call out the common symptom of scaling nodes into an AZ not covered by the VPC endpoint Related: elastic/support-tech-lead#1773 SDH ref: elastic/sdh-control-plane#11613
🔍 Preview links for changed docs |
Vale Linting ResultsSummary: 1 warning found
|
| File | Line | Rule | Message |
|---|---|---|---|
| deploy-manage/security/private-connectivity-aws.md | 147 | Elastic.Spelling | 'underutilize' is a possible misspelling. |
The Vale linter checks documentation changes against the Elastic Docs style guide.
To use Vale locally or report issues, refer to Elastic style guide for Vale.
cargious
left a comment
There was a problem hiding this comment.
@alxchalkias please review this as I remember we have produced a PM-like doc with the PrivateLink AZ-overlapping aspects.
@dnraitzyk @kunisen for a deep dive on PrivateLink flows you can read here. I believe the doc that Alex will provide is covering in high level what we need.
I agree with adding some Important flags and possibly add some more cases/examples, but we should not expose internal info for the NLB -> Elastic proxies hop.
|
Thank you @cargious for the review! I made an update based on your comment. My bad for the confusion, as well as the back and forth.
Could you please have a second check? 🙏 Screenshot after update:
|
|
Fine to make this more visible, I've only added one other minor edit. @shainaraskas, maybe you want to have a look at this yourself. While we have made the initial docs improvements many months ago, it seems like the information is not clear/visible enough to all customers, and Support still gets cases because of the traffic imbalance. |
Co-authored-by: Alex Chalkias <34575586+alxchalkias@users.noreply.github.com>
|
Thank you @alxchalkias !! One more thing - out side of this PR - while I was looking at the page, I found there are some discrepancies in between the lines [1]
[2]
I guess this happened during our doc internal (v3) migration. I will leave it to doc team's capable hands to decide how we handle this. |
cargious
left a comment
There was a problem hiding this comment.
I am approving the PR, though there are some docs-build suggestions that you might want to take into account.
Thank you again for making our docs better, it is really important for the customers and the load Elastic teams have when supporting the customers!
shainaraskas
left a comment
There was a problem hiding this comment.
small style edits. will open an issue for the other things you highlighted. this is not a result of the migration, but instead a long page being updated over time.
|
Approved. Sorry, I had not seen the notification |
Co-authored-by: shainaraskas <58563081+shainaraskas@users.noreply.github.com>
|
Thank you all for the help and review! I will merge. |
|
operational (ops) update - adding ech and serverless label (this applies to both ECH and serverless) |



Summary
Clarify the importance of ensuring user's VPC endpoint is in all availability zones supported by Elastic Cloud
Motivation
Per this internal ticket and this ticket
🤖 Partially Generated with Claude Code