[ixpmanager] 6.4.0 Upgrade issue?

[Moritz Frenzel]

Hi all,

just chiming in; we (Stuttgart-IX) are also seeing similar behavior;
we've seen a significant uptake in memory and CPU since the upgrade
(completed 20240525 ~21:00), see following link (expires in 7 days):
https://snapshots.raintank.io/dashboard/snapshot/IYGPyvNxQPAmZtP0hgrDZkONxhAsNT6I?orgId=0

We're also seeing the exact same PHP warnings.

Most noticeably, the P2P graphs for us are extremely slow, taking
~2min to render all ~30 peers.

Following along with the troubleshooting steps you requested below:

On Mon, May 27, 2024 at 5:18 PM Barry O'Donovan (INEX) via ixpmanager
<ixpmanager at inex.ie> wrote:
> SELECT   TABLE_NAME AS `Table`,
>        ROUND((DATA_LENGTH + INDEX_LENGTH) / 1024 / 1024) AS `Size (MB)`
>      FROM   information_schema.TABLES
>      WHERE
>          TABLE_SCHEMA = "ixpmanager" AND TABLE_NAME LIKE 'telescope%'
>    ORDER BY   (DATA_LENGTH + INDEX_LENGTH) DESC;
>
> changing "ixpmanager" to your database name as appropriate.

+------------------------+-----------+
| Table                  | Size (MB) |
+------------------------+-----------+
| telescope_entries      |       694 |
| telescope_entries_tags |         3 |
| telescope_monitoring   |         0 |
+------------------------+-----------+

>
>
> Once you've done that, and if big, you can also run the following in the
> database to clear it out:

not sure if the above is "big", I've still executed the commands.

>
> Also try adding the following to your .env:
>
> TELESCOPE_ENABLED=false

Done, is this supposed to remain in the config?

Since execution of the steps above, memory has freed up by ~700M and
CPU load has dropped by ~30% and the P2P graphs are fast again.