[X] My Assistant

A new system has been added to notify people if a H@H client they are running isn't working for some reason by automatically sending them a forum PM.

Firstly, if a client has been offline for more than 12 hours, a notification will be sent about this, advising to check on the client if this is unexpected.

Secondly, it will regularly attempt to connect to live clients to check for connectivity issues for incoming connections, and will notify if it is unable to connect to it for more than 12 hours.

Please let me know if you get any false alerts from this system, or if it fails to notify you when expected.

This is a fantastic addition. Now I won't only be getting emails, but a forum PM too when I forget to pay my invoices (IMG:[invalid] style_emoticons/default/heh.gif)

Thanks for the great feature (IMG:[invalid] style_emoticons/default/smile.gif)

Agreed, this seems quite helpful! (IMG:[invalid] style_emoticons/default/smile.gif)

Two related feature requests:
1. Could we schedule expected downtime? Every couple of years, I take clients down for a few days to migrate them to new servers. (This is minor, I can always just ignore the PMs.)

2. Could we get more detailed quality reporting? Right now there's just a single number, and it doesn't always make it easy to identify what the problem is. (E.g., for a long time, all my clients were typically at 10000 quality, but for the past 1-2 years, they've fluctuated more between 7000-9500 and I don't know why.)

Thankyou so much for continue supporting this beautiful site

Every day is a good day if I can be here (IMG:[invalid] style_emoticons/default/biggrin.gif)

1. Could we schedule expected downtime? Every couple of years, I take clients down for a few days to migrate them to new servers. (This is minor, I can always just ignore the PMs.)

It'll only send you one PM for any given downtime no matter how long it lasts, so yeah, that's probably not necessary.

2. Could we get more detailed quality reporting? Right now there's just a single number, and it doesn't always make it easy to identify what the problem is. (E.g., for a long time, all my clients were typically at 10000 quality, but for the past 1-2 years, they've fluctuated more between 7000-9500 and I don't know why.)

While there definitely is a lot more instrumentation data available than what is displayed, I'm not sure how useful it would be to show any if it, and having the system try to use it to draw any conclusions about the cause for any issues outside of the basic "quality" (i.e. it's performing well) and "trust" (i.e. it's performing correctly) would be pretty hard to do.

I did consider adding some fancy graphs and stuff at some point, but it would probably just show a high inverse correlation between traffic and quality, so that also probably wouldn't be actually useful for much of anything.

Is there's a option that we get the alert message through MoogleMail? I play HV much more often than I check the forum, so messages delivered through MM can be noticed earlier.

Good News: H@H 1.6 seems to have successfully delivered in-client pop-up warnings for clients experiencing high deviation in system clock time (I woke up to one such after about three weeks of mostly uninterrupted uptime).

Bad News: This was my first noted strike in-client but apparently my third noted via PMs (two messages delivered as "H@H Client 1259 is Offline", one as "H@H Client 1259 is Unreachable") and the client itself is throwing ERR_IDENT_REVOKED on startup again (known and readily addressed issue as seen in this thread and roughly a third of the Help Desk new threads).

Seems counterintuitive for the PM notification to say "restart the client in the next six months" if it's relatively quick to auto-revoke after racking up inactivity "strikes" based on either system clock deviation or forced-while-client-running outages (i.e. ISPs supplying you with routers/fiber that always want to drop off for ~90 seconds the second Peak Time in your area starts, or you make 2^n queries in your browser, or whatever local problem with your otherwise optimal connection can't be solved with user-side solutions).

This post has been edited by Andy Waltfeld: Mar 29 2021, 17:03

I confirm the system is working.
The Java VM crashed again (2nd time it happen, but this time I know it's the Java VM from previous experience), apparently 13 hours ago, so this morning I saw the message.
Convenient...though it would be more convenient if that stupid Java didn't crash in the first place (IMG:[invalid] style_emoticons/default/laugh.gif)