i see in monitoring that perf data is only showing up every 15mins. is there live per data that can be viewed?
Hello @gcp888 ,Welcome on Google Cloud Community.
For resources like SQL or GCE there is no possibility to change this default timeframe. You can see live data for Logs, by using "Stream logs" feature.
--
cheers,
DamianS
LinkedIn medium.com Cloudskillsboost
Hi GCP888,
I agree with @DamianS’ answer. Stream logs let you view log entries in real time in the log explorer, and there isn't a built-in real-time view of performance data in Google Cloud Monitoring.
However, you can use the Alerting Policy Service for additional performance data monitoring for your resources. The alerting policy can monitor time-series data stored by monitoring or logs stored by cloud logging. When that data meets the alerting policy condition, monitoring creates an incident and sends the notifications.
Here are the steps on how to create alerting policy using the Google Cloud Console:
In addition to creating alerting policies, you can use integrations and recommended alerting policies. Recommended alerting policies are helpful when you've deployed a new service and want to alert on important metrics.
I hope this helps!
Hi
So we have requirment to monitor the environment live. with live charts on GCE instances. SO google cloud monitoring does not have an option for this? so i will have to use something like grafana?
Hi gcp888,
Yes, that's correct. Unfortunately, Google Cloud monitoring does not have an option to monitor the environment in real-time or live. You can consider third party integrations like Grafana and Prometheus.
is there documentation anywhere saying live monitoring is not possible on gce?
We might need some clarification on what you mean by 'live' monitoring. For example, if you want to get metrics on a dashboard that are updated constantly, this could be seen as a 'live' dashboard and is entirely possible, see:
https://cloud.google.com/monitoring/charts/predefined-dashboards#disable-autorefresh
In terms of the 'age' of the data, arguably no data is ever 100% live, because it takes time to scrape and aggregate the metric samples, there will always be latency and a gap between samples, otherwise if you are polling a system constantly for data you risk the act of your own monitoring overloading a system.
Frequency of data polling varies by metric, if we look at a basic CPU usage of an instance, this is gathered every 60s, however there can be a delay for that to reach a monitoring dashboard due to aggregation and processing (as you can imagine the system is dealing with a vast volume of data and this takes time to process). For CPU usage this is listed as "up to 240 seconds". See:
https://cloud.google.com/monitoring/api/v3/latency-n-retention#latency
And:
So to summarise, for metrics like CPU usage, you can have a live dashboard that is streaming data with a resolution of 60s samples with a latency of up to 240s.
Hope that helps,
Alex