Grafna show strange result for influxDB

[s009255]

Hello!
I’m using grafana with influxDB 1.6.4.
I get strange results from some InfluxDB queries with group by tag.

SELECT mean(“cpu_utilization_average”) FROM “cloudwatch_aws_rds”
WHERE (“service” = ‘sge-cfz’
) AND $timeFilter GROUP BY time(1m), “service”, “db_instance_identifier” fill(previous)

it is result table panel of this query.
i think 11:55:00 result is not exist beacause time range is 11:55:01 ~ 12:00:01
but, 3 of 4 columns have results.

So, I use another query. it has a specific “db_instance_identifier” name.
SELECT mean(“cpu_utilization_average”) FROM “cloudwatch_aws_rds”
WHERE ( “service” = ‘sge-cfz’
and “db_instance_identifier”=‘live-cfc-p-rds-01-ap-southeast-1b’) AND $timeFilter GROUP BY time(1m), “service”, “db_instance_identifier” fill(previous)

it has different result!!
they have same time range but different result in 11:55:00.
Second query has no result in 11:55:00

In my opinion, This strange result is thought to occur if the same tag has a substring value like “One” and “OneTwo”.

In first image,
11:55:00 data of “live-cfc-p-rds-01-ap-southeast-1b” is same with last data of “live-cfc-p-rds-01”
i check it in different time range.

I need Grafana to show a correct result of this situation.

[jangaraj]

I’m not sure what is wrong, but it looks like you are surprised that you have data which you didn’t record → are you really sure that you want to use fill(previous)?

Check InfluxDB doc and you will understand what is that it doing - https://docs.influxdata.com/influxdb/v1/query_language/explore-data/

full(previous) - Reports the value from the previous time interval for time intervals with no data.

[s009255]

@jangaraj
Thank you for your reply.
It doesn’t have a problem without fill(previous).

But i think there is still something strange.
My influxdata is about ten minutes late.
I wanted my grafana graph to look more comfortable.
So i used fill(previous) because i usually look at the graph with now() - 30m.

query is SELECT mean(“cpu_utilization_average”) FROM “cloudwatch_aws_rds” WHERE (“service” = ‘sge-cfz’) AND $timeFilter GROUP BY time(1m), “service”, “db_instance_identifier” fill(previous)

Obviously, the time after which there’s no value is filled the last value present.
But it is strange in 11:44:00 data.
If i use fill(none) instead of fill(previous), 11:44:00 has no data.
But with previous 3 of 4 have strange value. it is not a previous time value of each.

Real value is shown above.

Is it normal to show like that?