Tier-3 gives Error

Help
,
1

My netdata Agent v1.38.0-136-nightly
Debian bullseye
is configured for Tier 4
But it seems I have overdone it?
in netdata.log:

grep DBENGINE error.log|grep ERROR
2023-02-23 07:17:02: netdata ERROR : MAIN : DBENGINE on 'mydomain.de': dbengine tier 3 gives aggregation of more than 65535 points of tier 0. Disabling tiers above 3 (errno 17, File exists)

I surelly oversized my config (more than 65535 )

   storage tiers = 4

    # Tier 0, per second data for a week
    dbengine multihost disk space MB = 1100

    # Tier 1, per minute data for a month
    dbengine tier 1 multihost disk space MB = 1330

    # Tier 2, per hour data for a year
    dbengine tier 2 multihost disk space MB = 1000
        # Tier 3
    dbengine tier 3 multihost disk space MB = 1000

should -i simply reduce the space MB to let’s say 500?

Disk space is not relevant for me, there is plenty enough.

2

Thanks @Bernd, apologies for being a few days late seeing this - let me see who can help us.

@Manolis_Vasilakis have you seen anything like this before?

3

Quick answer: stick to the 3 tiers for now, they will be more than adequate. You can increase the storage a bit and you’ll get many years of data stored.

The additional tier should still be working, but you’ll need to manually tune at least the default

dbengine tier 3 update every iterations

Get netdata.conf to see that these are set to 60 by default for every tier, to have a data point per second, minute, hour. A single data point per 60 hours isn’t very useful, but also runs into the following limitation:

Multi-Tier database backend for long term metrics storage by stelfrag · Pull Request #13263 · netdata/netdata · GitHub introduced a limit of 65535 points

Instead of storing packed numbers in it, we store unpacked all the values as a single number 16 bytes long (3x float + 2x uint16). So, the maximum number of points that can be aggregated is 65535 to any tier.

606060=216,000, so much more than 65535.

I will update the documentation

4

netdata/change-metrics-storage.md at master · netdata/netdata · GitHub has been updated, the change will make it to the nightly learn shortly.

5

Thank you guys for your updates. So, if I go back to 3 tiers, what a timeframe I do have for the History?

Every second, Netdata collects 12.445 metrics on 
mydomain.de, presents them in 5.920 charts and monitors them with 1.035 alarms.

The Base of compute this by myself is not really clear to me, I fear.
Nevertheless, You do do a great job here, A huge Thank You