docs: update onboarding steps for Java v3 client#6994
Conversation
wdoconnell
left a comment
There was a problem hiding this comment.
Hi @karel-rehor - thank you, I had a few observations here:
- This variant of the onboarding flow is behind the
ioxOnboardingfeature flag, so it won't be shown for any users with that flag disabled; the Flux onboarding will otherwise be shown. That flag is enabled generally, but I will check with others if that is the right approach here. - This makes the Java onboarding examples a good amount longer - I am not sure if that is desirable unless this is also intended as a bug fix or for some other purpose. Can you speak to the motivation behind the work?
- Not absolutely necessary in all cases, but for a PR like this that has visual impact on the flow, I usually find it is helpful for the review to include a video illustrating what was changed.
| const query = ` String sql = "SELECT * " + | ||
| "FROM 'census' " + | ||
| "WHERE time >= now() - interval '5 minutes' " + | ||
| "AND ('$species1' IS NOT NULL OR '$species2' IS NOT NULL) order by time asc"; |
There was a problem hiding this comment.
Are we updating this query due to a failure with the test dataset, or is this simply designed to illustrate a different principle?
There was a problem hiding this comment.
Query parameters are a feature added in the 0.6.0 client library release from March of last year.
The text above the snippet briefly introduces them: "The following code replaces the fixed values of "bees" and "ants" with the parameters $species1 and $species2."
We thought users would like to be made aware of this feature. A number of users requested it and are now using it.
| <p> | ||
| The query transport makes use of Apache Arrow Flight to shorten | ||
| processing time. When executing queries Arrow needs access to internal | ||
| JVM resources. This means setting the following JVM argument:{' '} |
There was a problem hiding this comment.
I might correct "means" to "This requires setting the following JVM argument"
| </p> | ||
| <h2>Java</h2> | ||
| <p> | ||
| With straightforward Java this can be done with an environment variable: |
There was a problem hiding this comment.
I don't think we should qualify whether the script is 'straightforward' since it should speak for itself -- "This can be done with an environment variable" should be enough.
|
Remocal failure in CI looks transient, so I am re-running the tests. |
|
Hi @wdoconnell
My understanding is that we primarily want these changes visible to cloud users and in the cloud interface, which, if I understand correctly, should have
The influxdb3 client libraries have evolved considerably since these onboarding steps were first written. Since the beginning of December client libraries are progressively being released to the first major version 1.0 (and in the case of go 2.0) because of breaking changes. The new onboarding steps illustrate the following features added in the past last year:
I also though it would be helpful to organize test data into a POJO as it is a common Java idiom,
OK, I thought that I would need to take time to compress the screencast file, as in the past I had encountered a 10MB limit, but it seems this can be added. This shows only the workflow without pasting snippets to an IDE. Screencast.from.08-01-25.10.57.08.webm |
wdoconnell
left a comment
There was a problem hiding this comment.
@karel-rehor: This looks good now. Thanks for your revisions, and sorry that it took me a bit to get back to another review.
With respect to this point:
This variant of the onboarding flow is behind the ioxOnboarding feature flag, so it won't be shown for any users with that flag disabled; the Flux onboarding will otherwise be shown. That flag is enabled generally, but I will check with others if that is the right approach here.
I think it's actually undesirable to have this controlled by flag, as opposed to the isOrgIOx selector. I discussed this with Daniel, and either he or I will put up a PR making that adjustment.
|
@karel-rehor Could you please re-run the pipeline for this? It looks like the past failure was flake. If not, please ping me and I'll diagnose the CI failure. |
|
@wdoconnell Hi, I've rerun the monitor-ci-tests and they are failing on the workflow dependency "remocal-smoke-test" and then In this job the step "install kubectl binary" is failing. It downloads the binary from (https://storage.googleapis.com/kubernetes-release/release/v1.15.0/bin/linux/amd64/kubectl) and then runs a sha256 check sum of the downloaded file. The download fails, (most likely because of an issue with locating certificates - see first line below) and the kubectl executable can then not be found on the system. Unfortunately I don't have permissions to rerun this job with SSH to investigate further. |
|
Hi @karel-rehor - I traced this, and we have someone fixing this internally. Since this is public I am reserving the details for slack. |
|
Hi @karel-rehor : You should be good to rebase and proceed with trying to merge, as CI should be fixed. |
|
Hi @karel-rehor - checking in to see if you are good to rebase and merge this - thank you. |
64572be to
7e83816
Compare
|
@wdoconnell OK, This has been rebased. Do you want to rebase this or should I handle it? |
You're all set now @karel-rehor -- feel free to merge. I'll keep an eye out if CI fails, but it should be clear now. |
|
@karel-rehor Looks like transient flake to me - I am re-merging. |
* docs: (WIP) updates onboarding steps for the influxdb3-java client. * chore: dry run onboarding java. Add missing semi-colon. Minor edit. * chore: run prettier:fix * docs: update wording of query requirements for Arrow. * chore: rerun linter * chore: retrigger checks
Updates the onboarding steps for the Java v3 client.
Introduces new library features.
Checklist
Authors and Reviewer(s), please verify the following: