Fix Couchbase regression

[akrantz01]

Required for all PRs:

• Updated associated README.md.
• Wrote appropriate unit tests.

resolves #9416

Fixed an out-of-bounds panic that occurred when one of the bucket stats fields was missing in the response from Couchbase. A check to see if the array is nil was added to all bucket stats fields.

[akrantz01]

@Levarix would you be willing to test once the binaries are built?

[telegraf-tiger]

Looks like new artifacts were built from this PR. Get them here!

Artifact URLs

• darwin_amd64.tar.gz

• windows_amd64.zip

• windows_i386.zip

• s390x.rpm

• linux_s390x.tar.gz

• s390x.deb

• armel.rpm

• linux_armel.tar.gz

• armel.deb

• ppc64le.rpm

• linux_ppc64le.tar.gz

• ppc64el.deb

• i386.rpm

• freebsd_i386.tar.gz

• linux_i386.tar.gz

• i386.deb

• x86_64.rpm

• freebsd_amd64.tar.gz

• linux_amd64.tar.gz

• amd64.deb

• linux_mips.tar.gz

• mips.deb

• armv6hl.rpm

• freebsd_armv7.tar.gz

• linux_armhf.tar.gz

• armhf.deb

• linux_mipsel.tar.gz

• mipsel.deb

• aarch64.rpm

• linux_arm64.tar.gz

• arm64.deb

• static_linux_amd64.tar.gz

[srebhan]

Looks good to me.

[Levarix]

@akrantz01 I've tested this, and now we have another huge problem. I applied a new version of telegraf in a production environment and telegraf processes consumes a lot of memory and don't purge it. Memory usage is still increasing until instance stopped work. It happens only with telegraf with couchbase plugin

In the second screenshot, you can see memory usage increasing and after downgrade to 1.18.3 back to normal stable consumption

[srebhan]

@Levarix can you please open an issue for this?

@akrantz01 the problem might be in

func (cb *Couchbase) queryDetailedBucketStats(server, bucket string, bucketStats *BucketStats) error {
	// Set up an HTTP request to get the complete set of bucket stats.
	req, err := http.NewRequest("GET", server+"/pools/default/buckets/"+bucket+"/stats?", nil)
	if err != nil {
		return err
	}

	r, err := client.Do(req)
	if err != nil {
		return err
	}

	defer r.Body.Close()

	return json.NewDecoder(r.Body).Decode(bucketStats)
}

as the final json.NewDecoder(r.Body).Decode(bucketStats) will only decode the first valid JSON object, so if therer are multiple JSONs sent, the decoder will not read all off r.Body and thus might leak memory. Not sure though but maybe it gives you a hint?

[Levarix]

Created @srebhan #9495