Timeout calling getAccountFunds

[bnl]

Hi!
Every 30 seconds I check my balance with
getAccountFunds, to see if it is ok to place a bet.
I'd like to keep the exposed part below a set threshold.

But lately, I get a lot of timeouts for that call.
I called the api ~6000 times today and yesterday,
and in 35 case (almost 0.6 %)
I get a timeout.
The post method has a timeout of 30 seconds

Is there something going on at Betfair?
I have not noticed this before.

2015-06-16 12:39:23.660 RPC.Get_JSON_Reply : posting: {"method":"AccountAPING/v1.0/getAccountFunds", "jsonrpc":"2.0", "id":15, "params":{}}
2015-06-16 12:39:53.739 RPC.Get_JSON_Reply : Post Timeout -> Give up!

Times are in CET

As a consequence, I terminate the bot, and restart it later.
This could be improved, but I wonder what is the best thing
to do here?

To logout/login?
To retry ?
To increase Timeout on POST? - to 60s ? 120s?

Commennts anyone ?

/Björn

[JayBee]

Yes, I have noticed that too. A couple of times a day getAccountFunds fails. You can see the failure, server side, with Fiddler. It can take minutes for the response so I stopped calling the method.

I decided to bite the bullet and set up an automated login with a digital certificate. Now, when the program attempts to fall over, it will be directed to login again, get a new SSOID and then continue. I still have two-part authentication in case anyone tries to crack the login via the website. The protection of the self-sign certificate is my responsibility.

All these connections drops are making a mess of my data collection so something had to be done about it. As usual, client side.

[betdynamics]

It's been appalling for the past few days, but seems even worse today :-(

Might be just coincidence, but it does seem to be worse since the recent work that was carried out on the network.

[Lopiner]

I have the timeouts also on getAccountFunds since the beginning of this API

[bnl]

Ok - good to know I am not alone.
Let's hope they will fix it soon.

I wonder - it seems that most of the times
betfair deploys a new version of the api,
something breaks.

Or is it me missing all the successful deploys,
since I do not notice anything bad ?
I hope it is the latter.

/Björn

[BetfairDeveloperProgram]

Hi Björn

These issues have been reported to our dev team. However, I would recommend that you handle failure by retrying the requests as we cannot guarantee a 0% failure rate.

Thanks

Neil

[BetfairDeveloperProgram]

We believe we've identified the issue which is causing the getAccountFunds timeouts and this requires an update to the underlying API framework - the date for this update is TBC.

In the meantime, we have made a temporary fix which we believe will reduce the overall impact of this issue.

Thanks

Neil

[bnl]

We believe we've identified the issue which is causing the getAccountFunds timeouts and this requires an update to the underlying API framework - the date for this update is TBC.

In the meantime, we have made a temporary fix which we believe will reduce the overall impact of this issue.

Thanks

Neil

Good to hear , thanks
/Björn

[bnl]

Hi Björn

These issues have been reported to our dev team. However, I would recommend that you handle failure by retrying the requests as we cannot guarantee a 0% failure rate.

Thanks

Neil

Hmm, my code doing the communicating does not really know what type of message it sends, same code sends these kind of messages as the bet placement messages. And I do not want to risk they a bet actually gets through, and I then resubmit it, thus betting twice.

I can live with these failures, I just do not want them as often as now. The bot is auto-restarted, so I just miss one bet.

/Björn