I tried to visualize some histogram metrics in a dashboard to show call durations and found that for no histogram metric I was able to produce such thing using the Query Builder. I always get an error. Do I do something wrong here?

No, this looks fine. If you still get error, please share the logs of query-service when see this error.

I checked the query-service logs and found those errors:

signoz-query-service  | {"level":"ERROR","timestamp":"2024-10-28T06:54:28.621Z","caller":"clickhouseReader/reader.go:4446","msg":"error while reading result","error":"code: 302, message: Child process was exited with return code 88: while executing 'FUNCTION histogramQuantile(arrayMap(lambda(tuple(x), toFloat64(x)), groupArray(le)) :: 5, groupArray(value) :: 2, 0.9 :: 4) -> histogramQuantile(arrayMap(lambda(tuple(x), toFloat64(x)), groupArray(le)), groupArray(value), 0.9) Float64 : 1'","stacktrace":"<http://go.signoz.io/signoz/pkg/query-service/app/clickhouseReader.getPersonalisedError|go.signoz.io/signoz/pkg/query-service/app/clickhouseReader.getPersonalisedError>\n\t/home/runner/work/signoz/signoz/pkg/query-service/app/clickhouseReader/reader.go:4446\ngo.signoz.io/signoz/pkg/query-service/app/clickhouseReader.readRowsForTimeSeriesResult\n\t/home/runner/work/signoz/signoz/pkg/query-service/app/clickhouseReader/reader.go:4300\ngo.signoz.io/signoz/pkg/query-service/app/clickhouseReader.(*ClickHouseReader).GetTimeSeriesResultV3\n\t/home/runner/work/signoz/signoz/pkg/query-service/app/clickhouseReader/reader.go:4383\ngo.signoz.io/signoz/pkg/query-service/app/querier/v2.(*querier).execClickHouseQuery\n\t/home/runner/work/signoz/signoz/pkg/query-service/app/querier/v2/querier.go:107\ngo.signoz.io/signoz/pkg/query-service/app/querier/v2.(*querier).runBuilderQuery\n\t/home/runner/work/signoz/signoz/pkg/query-service/app/querier/v2/helper.go:243"}
signoz-query-service  | {"level":"INFO","timestamp":"2024-10-28T06:54:28.621Z","caller":"utils/time.go:17","msg":"Elapsed time","func_name":"GetTimeSeriesResultV3","duration":648,"path":"/dashboard/dd07c9b8-302b-4e98-a008-c0993dbbb79e/new","query":"SELECT ts, histogramQuantile(arrayMap(x -> toFloat64(x), groupArray(le)), groupArray(value), 0.900) as value FROM (SELECT le, toStartOfInterval(toDateTime(intDiv(unix_milli, 1000)), INTERVAL 240 SECOND) as ts, sum(value)/240 as value FROM signoz_metrics.distributed_samples_v4 INNER JOIN (SELECT DISTINCT JSONExtractString(labels, 'le') as le, fingerprint FROM signoz_metrics.time_series_v4_1day WHERE metric_name = 'signoz_latency_bucket' AND temporality = 'Delta' AND unix_milli >= 1729987200000 AND unix_milli < 1730098440000) as filtered_time_series USING fingerprint WHERE metric_name = 'signoz_latency_bucket' AND unix_milli >= 1730011920000 AND unix_milli < 1730098440000 GROUP BY le, ts ORDER BY le ASC, ts ASC) GROUP BY ts ORDER BYts ASC","dashboardID":"dd07c9b8-302b-4e98-a008-c0993dbbb79e","alertID":"","source":"dashboards","client":"browser","viewName":"","servicesTab":""}

Is this some kind of timeout? I did just limit the query to 1 day and it timed out after 648ms?

No, it's not a timeout. We use UDF to calculate the final latency. The UDF exited with an error. What is your CPU arch?

We are using an arm-based EC2 instance (c6g.8xlarge).

I think that might the problem. The UDF might be breaking on arm based machines.

So should I open a Github issue about it?

Yes