On Friday, 25 January 2019 at 16:13:24 UTC, Vladimir Panteleev wrote:
> On Friday, 25 January 2019 at 15:33:37 UTC, Steven Schveighoffer wrote:
>> Just this morning I clicked on a link in a post (use Thunderbird normally, so I'm not a regular user of the forum), and it was "down for maintenance" 2x, then finally came up.
>
> Would you happen to have a more exact time?
>
> Here are the holes in today's log (UTC):
>
> 2019-01-25 07:02:49 - 2 minutes, 30 secs - daily backup
> 2019-01-25 14:53:54 - 4 minutes, 42 secs - unusual lag spike
> 2019-01-25 15:02:36 - 1 minute, 38 secs - unusual lag spike
>
> I still need to figure out those lag spikes (single requests / database queries that take much, much longer than they should), but knowing if it's one of those would allow me to know if the problem you saw was there or elsewhere (such as the reverse proxy).

Proxy Error

The proxy server received an invalid response from an upstream server.
The proxy server could not handle the request GET /first-unread/ltfxzjpiuaryxqjcbzrj@forum.dlang.org.

Reason: Error reading from remote server

Apache/2.4.18 (Ubuntu) Server at forum.dlang.org Port 443

Just as I entered this thread to read the two new messages :)

On 1/25/19 11:13 AM, Vladimir Panteleev wrote:
> On Friday, 25 January 2019 at 15:33:37 UTC, Steven Schveighoffer wrote:
>> Just this morning I clicked on a link in a post (use Thunderbird normally, so I'm not a regular user of the forum), and it was "down for maintenance" 2x, then finally came up.
> 
> Would you happen to have a more exact time?

According to my chrome log it was at 9:55 am which corresponds to 14:55 UTC.

-Steve

On Friday, 25 January 2019 at 12:41:35 UTC, Vladimir Panteleev wrote:
> On Monday, 21 January 2019 at 07:52:49 UTC, Vladimir Panteleev wrote:
>> I've been trying to find a more permanent solution to the performance problems plaguing the forum (and other D services hosted on the same server), and trying several things to make it work with the current hardware.
>
> As of a few days ago, the forum's database is now on a dedicated partition. (Even though it's on the same physical disk, the dedicated partition allows bypassing the sync queue from the rest of the system.)
>
> Performance should now be much be much better, and according to my observations, it has improved.
>
> What is your experience?

Yesterday was clearly better but this morning i just got the "service unavailable" page.