Chess.com server status

Sort:
Oldest
Homer1990

Hello! As a fellow server admin (just an enthusiast, not chess.com level, but with about 16 years of experience) I am proposing that a general, not detailed, server status tab be added, either as an optional "gadget" or as a mandatory footnote-style section, because when the server(s, CDNs?) suffer an issue, many games are lost or cancelled. I lost a number of games in the last half-hour due to spontaneous cancellation, that was caused either by maintenance work or by some infrastructural issue. Players should know when they should avoid playing, because time and points are lost when one does not know if they can play or not.

Homer1990

The server "going down" is one of many possible issues. You need to take into account the the "server" is a collection (in this scale) of many distributed copies of a number of "servers". And I put it in quotes because there is a vast difference between a physical server (the computer) and what we mean by server, which is server, the program that provides a service.

A site like chess.com needs a great number of physical machines running "copies" of those services and databases, with an number of inner machines that update the final status of the site globally (ie, for all machines running those core services). That is the site core.

The "decision": "which machine (or, most likely, set of machines) will provide service Y to user X" is made by using a number of intermediary services on different machines, colloquially known as load balancing/load balancers.

At (the very) least for those that are somewhat more adept at such concerns, a set of basic information (regarding one's location and connection, of course), specifically about the state of performance (basic latency, load time server-side, state of mirrors etc.) must be provided along with an estimate from the site itself (eg. "there is very high usage of the database servers, games are dropped" or "CDNs in <user_country> are experiencing technical issues" etc.).

Martin_Stahl
Homer1990 wrote:

Hello! As a fellow server admin (just an enthusiast, not chess.com level, but with about 16 years of experience) I am proposing that a general, not detailed, server status tab be added, either as an optional "gadget" or as a mandatory footnote-style section, because when the server(s, CDNs?) suffer an issue, many games are lost or cancelled. I lost a number of games in the last half-hour due to spontaneous cancellation, that was caused either by maintenance work or by some infrastructural issue. Players should know when they should avoid playing, because time and points are lost when one does not know if they can play or not.

 

There was a planned server restart and when thatt is scheduled, there is banner in live with the information about. There was a second restart after the first, to fix an issue. Both restart had the banner.

 

Potential CDN problems don't impact the Live server process and as of right now, there is only one Live server when playing rated games. 

Homer1990
Martin_Stahl έγραψε:
Homer1990 wrote:

Hello! As a fellow server admin (just an enthusiast, not chess.com level, but with about 16 years of experience) I am proposing that a general, not detailed, server status tab be added, either as an optional "gadget" or as a mandatory footnote-style section, because when the server(s, CDNs?) suffer an issue, many games are lost or cancelled. I lost a number of games in the last half-hour due to spontaneous cancellation, that was caused either by maintenance work or by some infrastructural issue. Players should know when they should avoid playing, because time and points are lost when one does not know if they can play or not.

 

There was a planned server restart and when thatt is scheduled, there is banner in live with the information about. There was a second restart after the first, to fix an issue. Both restart had the banner.

 

Potential CDN problems don't impact the Live server process and as of right now, there is only one Live server when playing rated games. 

Hello! I had no such banner. I was live for hours before what you now say was a restart.

Martin_Stahl

While I can't definitively say it was showing for you, in the times I've been on when one was scheduled, I've always seen it.

 

It looks something like the following:

Homer1990
Martin_Stahl έγραψε:

While I can't definitively say it was showing for you, in the times I've been on when one was scheduled, I've always seen it.

 

It looks something like the following:

 

Hello! I know, I've seen it too. Not today, though.

My main point is that certain metrics about site performance should be made available on a permanent basis. 

Brooksvillechess

Especially in light of current events, this seems like a very good idea. It would be very useful to be able to anticipate issues a bit more accurately. 

Martin_Stahl
Brooksvillechess wrote:

Especially in light of current events, this seems like a very good idea. It would be very useful to be able to anticipate issues a bit more accurately. 

 

Most metrics that would make sense to people, would give any information to anticipate potential issues. Maybe some very basic load information could potentially give a clue, but higher loads don't automatically mean potential issues, depending on where the load is.

 

And anything automated would not be seen when the site is having issues, unless it's hosted completely separate from any other site resources.

 

Though, I agree, some kind of site status page, would be nice. Maybe some uptime report, scheduled restarts for Play, and recent issue summaries.

Brooksvillechess

Ok, thanks. 

Forums
Forum Legend
Following
New Comments
Locked Topic
Pinned Topic