It's been a couple weeks now that my chance for getting jobs took a noticeable plunge. Most of the jobs, even the small ones (< 1$) are gone the second I clicked them. I feel that this 'symptom' is pretty similar to back when the cache was not working. Of course it could just be my bad luck, but at this point I think there's something else here. Any one feel the same? Could someone from Gengo crosscheck this with the tech team?
Is cache not working - again?
Yes, same for me, I basically never get any of the jobs my RSS feed gives me notifications for.
In my case, it started since the early December. I honestly can't tell if it was the problem with the cache, or the fact that the volume of jobs was simply very low because of the holiday season. Nevertheless, it'd be nice if we have some kind of confirmation from Gengo regarding this matter.
Hi all - and thank you for bringing up this issue :) I've looked into it and our Engineering team has confirmed that nothing has changed in the system, and that the cache is still working properly. Like @Ryan mentions, depending on your language pair there may have been a slight decrease in job volume during the holiday period, but we'd expect that to pick back up once businesses resume their regular activities.
Hope this helps and, for any technical issues, please don't hesitate to email us at firstname.lastname@example.org.
I have noticed this too in my language pair. I think it began when Feeder was released, a few months prior (August?).
Hello, I absolutely agree with Val and Raymond and I add that, from my point of view, the problem is going from bad to worse. When asking email@example.com about the issue, everybody beats about the bush but nobody speaks clearly. Kind regards.
I don't know if it's the same thing you are referring to as "cache", but I did notice those past few months a problem that I hadn't encounter before. When I click on a job notification from my RSS Feeder (either Rumpel or RSS Feed Reader), it happens sometimes that the page gets stuck before I can reach the Dashboard. It just happened to me again, it took more than 20 seconds for me to reach the Dashboard, and by then the job was of course already taken. The URL always says the same thing "referral=rss" and at the bottom "waiting for gengo"...
I wrote to Support and was told that everything was fine on their end, it might be my connection, etc. but I'm sure it wasn't.
I understand that there might be fewer jobs at the moment and many translators as desperate as I am who click as soon as they see a job, but this "freeze" is getting extremely irritating at best, and has costs me many jobs already...
@Lara, even if there haven't been any technical changes, it seems to me that the server has simply become dead slow. This could be due to high usage and workloads or whatever tech problems and you should do something about it. Frankly, in the last few months I began to ignore any small assignments simply because it is not worth the wait. A year ago or so, when I clicked on the link in my RSS window after a job popped up, the Gengo page with the job opened in my browser and I could either click on "start translating" or I was told that someone else was faster, okay, that's life. Nowadays, when I click on the RSS link, it takes ages even to open the Gengo page. It can take one or two full minutes, and I can really make better use of my time than waiting for a $1.50 job that will be gone anyway when the page finally loads.
I also have seen the "freeze" @Xavier mentioned and it's an additional nuisance. You really need to do something about it. Your tech guys should be able to test this themselves with some sort of test account, and then find the cause and fix it. Gengo is about automation and smooth workflows, and this problem is really disrupting everything.
See also here and here
I wonder if the problem also depends on the location of server and translator.
I definitely have noticed severe delays from Germany at times.
@Lara - Thank you for checking and confirming with the tech team. Maybe it is my luck after all, although the recently lower volume of job in my pair doesn't help either.
Also, could you please address the slow load time of the workbench page, as you could see here the issue is affecting many of us and I agree with the others that Gengo need to do something about this one.
@Ballmar - This is precisely the reason why I went straight here instead of contacting the support, I almost feel that nowadays they aren't as 'open' as they were back in the old days...
@Xavier - I'm no tech guy, but I believe it is not the cache issue but rather the server inability to cope with the surge of translators trying to access the said page, and I also experiencing the same issue as yours from time to time.
I absolutely agree with @kvstegemann post, that they need to address this issue as it affects our ability to pick up the jobs and it has been going on for some time now. I noticed after the system maintenance on Christmas things worked great - pages load fast for a couple days, but after that it's back to slowing down again.
@Chris - I'm in Indonesia and am having the same delay from time to time, so I think it is a system-wide issue.
@Raymond et al. - Thank you for this. I understand the issue is affecting your ability to pick up jobs, and all feedback from this thread has been passed to our Engineering team. It's my understanding that they're currently investigating the cause for the slow loading times. I will update you when I receive more information.
I just submitted a ticket to support, but I thought I would ask here whether other translators experiencing the same problem with the workbench since a few days.
'Since a few days workbench has been very slow and often gets stuck when loading matches, requiring me to reload a new page with the same collection before continuing. When I turn off matches it works just fine, but this requires translating a lot of content again without using any (partial) matches present.'
@philip.lardot - I think the situation is the same as here.
Got a very quick response from the support team.
Thank you for writing in.
Our DEV team is aware of it and they suspect something recently deployed causes this issue and they are working to solve it.
Thank you for your patience in the meantime.'
Hi all - I was wondering if you guys have noticed an improvement in loading times?
The Engineering team has shared with me that, according to our metrics, the workbench should be much faster now, and that these improvements should have become evident after Jan 16th :)
Could you guys confirm?
Not sure. It still takes too long in my view. I also noticed lots of activity in the status bar of my browser, and this activity is not only coming from gengo.com, but also from doubleclick.net, googleadservices.com, yahoo.co.jp, yimg.jp, and some more that runs too fast through to read it. I think stuff like that should not be present on a page that is designed for work, it looks quite suspicious and it might increase loading times for users in certain parts of the world.
Maybe someone could try a good ad blocker and see if this improves loading times?
For me, it's still so slow that I've had to reject several [customer name has been edited out due to privacy reasons] jobs because I cannot work on them. It starts off slow but workable, but after the first five or six lines it essentially crashes.
While I'm not sure about the speed of the workbench, but it definitely seems to be crashing more than before. I'm losing my work multiple times per collection and [customer name has been edited out due to privacy reasons] jobs are basically impossible. I certainly haven't noticed any improvement I'm afraid.
Hello, Lara, and all of you,
Since Gengo’s database updating, one month ago (in my opinion, to emphasise this fact is important) the system doesn’t work well. I have repeatedly reported the issues mentioned on this forum, and others affecting me individually, via firstname.lastname@example.org.
The answer I got was that there was nothing wrong and that the system was working correctly, but that statement is not compatible with what really happens (?). These problems are not of minor importance, mostly because we are suffering them for more than a month. At least, it would be convenient for all of us to know if the problems have been already detected by the DEV team and, consequently, they are working to fix them.
Regarding my colleagues comments, I corroborate all the issues pointed out by them and I know too well what they are talking about. Currently, to complete most of collections means to go through hell, as it takes three times more time than it would take in normal circumstances: the page gets blocked, it gets stuck, it crashes all the time, when opening the matching glossary, the page gets stuck, etc, and that doesn’t happen from time to time but constantly. The described situation forces me to refresh the page once and again and, sometimes, this is not enough to retake translation, and I have to reset the computer.
Thank you for looking into the matter and for keeping us informed.
Unfortunately I have to agree with the translators above.
Working on [customer name has been edited out due to privacy reasons] jobs has become incredibly difficult when using any TM/matches. Every third or fourth job the whole workbench freezes because it can somehow not load the TM data. This requires the translator to refresh the browser or often even to open a new tab, while also losing the last jobs he/she did due to the workbench freezing and not saving.
The only way to work on these jobs currently is by turning off matches. In the case of [customer name has been edited out due to privacy reasons] however, where there is a great number of 95-99% matches that we can not use currently, this means translating a lot of content again instead of just making a minor modification to something from TM. Obviously this takes much more time and the actual reward for the translation has been decreased due to the TM match.
Thank you all so much for your feedback! I have passed your comments to our Engineering team and they will continue investigating so that we can fix this issue as soon as possible. In the meanwhile, please continue to comment here with any updates or changes in your experience. I will continue to follow this thread closely.
We apologize for the inconvenience and are hard at work to sort it out. Thank you for your patience!
I happen to be in the middle of translating a [customer name has been edited out due to privacy reasons] collection right now, and I am observing the exact problem that everyone here describes: TM matches load for the first few segments, and then the screen slows down and freezes. This has been happening consistently for the last three weeks (before and after the 16th of January), so it should be easy enough for the development team to reproduce! It just has to be a fairly large collection (the one I am working on is 42 jobs, which is a typical size for my pair; it froze after I translated less than a dozen). In fact, the only way to avoid seeing this is to hide the matches.
I am also wondering if Gengo is considering adjusting rewards in some way while this problem is being fixed.
I am just worried that this ends up becoming one of those issues that everyone was very vocal about for a while and then gave up on (like file preview), so I really appreciate that you are following up on it. Please don't give up, Lara!
Just to clarify for the engineering team, I think there are three main problems that people are experiencing:
1. Workbench loading slowly and sometimes crashing after receiving an RSS message of a job being available and following the link. (Seems to me that this has been fairly consistent in the last +/- year.)
2. Workbench freezing when loading TM matches, not always but generally within a few minutes. (A problem that arose somewhere in December?)
3. Workbench just being incredibly slow, especially with larger collections. When I have reached about half or 2/3 of a large collection, the workbench often becomes so slow that the letters I am typing are not actually displayed until a few seconds later. (Also a problem since somewhere in December)
Personally I think problem number one is the least of our worries, but the other two relatively new problems have made translating larger jobs very tedious. I would estimate that these jobs take at least 25% longer now than they did before. Combined with the fact that we pretty much have to work without TM and translate a lot of content for a reduced rate without being able to use the TM match, it is even worse.
As for point 3, if the situation can not be fixed in the short term, as seems to be the case, would it perhaps be possible to break up collections into smaller collections with which the workbench seems to have less trouble? (I think this goes mostly for [customer name has been edited out due to privacy reasons] collections)
I think 1. has improved recently, but I'm busy with other things these days, so I can't tell for sure. If it happens, it can be very frustrating, though.
I agree with philip.lardot about temporarily splitting [customer name has been edited out due to privacy reasons] collections up in much smaller units - perhaps having a single lot as a job by itself? That would at least mean we could work on them in the short term.
Although to list the problems we currently pointed out on this thread, and specially to clarify them is positive, to decide priority in fixing them is quite personal and, regarding all the underlined issues here, there are many people involved who are affected to different degrees by all of them, specially by point (1), therefore, the aim should be to fix them all, not only some of them on demand basis. And also, if the main subject of the (3) points deals with ‘workbench slow loading’ why then you conclude that point (1) is less important?
Among the affected by point (1), according to what I understand they meant: Raymond, Juliane, Val, Xavier, HQ, Kat.Lev and myself (I can stay days without being able to pick up a job, although I translate in two pairs with quite steady jobs volume), and I guess that this is due (among other unknown reasons) to workbench slow loading)). That said, you can imagine that for the mentioned translators and myself, to fix point (1) is imperative, even though we haven’t marked it as such.
I did not mean to downplay the importance of point number 1, obviously it should be of great importance to have the workbench load quickly and reliably for every translator.
As I stated in my previous post, the prioritization was personal for me and I imagine it probably applies to most other translators in higher volume language pairs, where 'catching' jobs is not the hardest part. Obviously this will be different for all translators though, no offense was meant.
I have encountered both problems, actually, it's just that there are no [customer name has been edited out due to privacy reasons] projects in my language pair (or I have never seen them) and it's been getting more and more difficult to grab the really large jobs. But I was able to work on some large projects for the moving company and also suffered from the incredibly slow loading times on the workbench. Since everything I typed took ages to appear on the screen, I would type on Notepad and then copy-paste.
I think what would work is if, instead of having the remaining units load on the same page, we could click "next" and have a new page open. This has always happened, really, maybe now with the TM it is worse. But the really massive collections have always slowed down as you progressed.
No, I didn’t feel offended at all because you did nothing offensive, just on the contrary, new points of view are always positive, I just meant that I understand that to priorise is not the aim of this thread, and also, I don’t think this is up to us to decide on this matter. Anyway, I understand that the aim of all of us is to have all the issues reported here, fixed as soon as possible.
I'm just dropping by here to say that I, too, am experiencing all of the problems listed above and that it certainly has not been getting any better lately; if anything, it may have actually even been getting worse.
Same issue here. It's extremely frustrating... Working on a collection no matter its size now takes twice as long (if not more).