The minute crown and the hour crown stopped working as a result, no ROZETKA, PRIVAT24 integrations work

Fixed, there was a crash.

[quote] Tyndyk Maxim Vadimovich Administrator wrote: Fixed, there was a crash. [/quote] Hi As far as I understood, many clients had this failure (later I saw 2-3 similar problems) If I understood correctly, then the boxes were updated and because of some bug it stopped working I have a question: 1. Can you implement some kind of mechanism that can catch such problems, since now you catch such problems with the help of clients (the problem is that such things break all integrations, and this can lead to bad consequences)? 2. Why did cron-minute and cron-hour not work, but cron-imap worked (it has some other mechanism)?

[quote] Kupriyan Vladislav Valerievich Baza.cn.ua / Integrator (FOP Kupriyan) wrote: 1. Can you implement some kind of mechanism that can catch such problems, since now you catch such problems with the help of clients (the problem is that such things break all integrations, and this can lead to bad consequences)? [/quote] support

You just understand that fixing after a while is not enough. All the same, there is time when the system was lying, and this can be critical for business. We need mechanisms that will instantly respond to such situations and work will be restored as soon as possible. A few hours off work is simply unacceptable. What if it's a 24/7 call center?

maybe you will make some kind of knopulina on the forum like "Alarm, there is a serious problem", which can be clicked ONLY once a month ONE user of the forum? Information on such a problem will be sent to whoever needs it via SMS, and will not hang on the forum for X hours. those. to be able to provide a quick response to the task

[quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: maybe you will make some kind of knopulina on the forum like "Alarm, there is a serious problem", which can be clicked ONLY once a month ONE user of the forum? Information on such a problem will be sent to whoever needs it via SMS, and will not hang on the forum for X hours. those. to be able to provide a quick response to the task [/quote] It seems to me that this is too complicated, and again, this is already a belated reaction that essentially relies on living people, you need to do some kind of functionality that will signal to the owner that something is wrong with the cron, like there is an action "Send email, if one of the system indicators has reached the specified load" do something similar so that the admin receives a signal, and then turns on the "Alarm" (in principle, the issue was resolved on the forum quickly) for an hour. I would like to receive an answer 2. Why did cron-minute and cron-hour not work, but cron-imap worked (it has some other mechanism)?

[quote] Kupriyan Vladislav Valerievich Baza.cn.ua / Integrator (FOP Kupriyan) wrote: [quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: maybe you will make some kind of knopulina on the forum like "Alarm, there is a serious problem", which can be clicked ONLY once a month ONE user of the forum? Information on such a problem will be sent to whoever needs it via SMS, and will not hang on the forum for X hours. those. to be able to provide a quick response to the task [/quote] It seems to me that this is too complicated, and again, this is already a belated reaction that essentially relies on living people, you need to do some kind of functionality that will signal to the owner that something is wrong with the cron, like there is an action "Send email, if one of the system indicators has reached the specified load" do something similar so that the admin receives a signal, and then turns on the "Alarm" (in principle, the issue was resolved on the forum quickly) for an hour. I would like to receive an answer 2. Why did cron-minute and cron-hour not work, but cron-imap worked (it has some other mechanism)? [/quote] [quote] Kupriyan Vladislav Valerievich Baza.cn.ua / Integrator (FOP Kupriyan) wrote: It seems to me that this is too complicated, and again, this is already a belated reaction that essentially relies on living people, you need to do some kind of functionality that will signal to the owner that something is wrong with the cron, like there is an action "Send email, if one of the system indicators has reached the specified load" do something similar so that the admin receives a signal, and then turns on the "Alarm" (in principle, the issue was resolved on the forum quickly) for an hour. [/quote] you describe an ideal picture, but the problem is deeper than just in crons, but in updates that are loaded without testing for live projects. I'll tell you how it was In the evening of the previous day, for one of the projects, a revision was unloaded, which I ordered, and requests from users immediately rained down that something was not right. The problem has been confirmed. I wrote to the developers about it, phoned support, but since it was at 5:40 pm, everyone was probably happily drinking tea and getting ready for bed. I understand that if improvements were uploaded to the project and this terminated the project, then either in this revision or in the general update package there is clearly a bug that will spread to other projects with the next update. and if I had the opportunity to press a similar button and the developers received express information about the problem, then I think this topic would not exist. and this situation can arise not only with crowns. they can screw up the starter, and the business process module ... yes, anything, and writing autotests for each module means paying for XXX hours of programming. If there was a readiness for this, then I think they would have already written. I actually suggested the simplest way to solve the problem in my opinion and the most fault-tolerant one. We are all sane people here who understand that if something is wrong with one box and this is not a feature of a particular box, then you need to poke "Alarm" and tell the developers about it until the rest get worse. For it is always more difficult to rake up the consequences of a fire than to put out an ember that has fallen out of the furnace.

[quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: which are loaded without testing on live projects [/quote] problem

[quote] Farkhshatov Rodion shine-bright.com.ua wrote: [quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: which are loaded without testing on live projects [/quote] problem [/quote] which you came up with here in the topic In fact, the reason is not that you are discussing here, but not even close, but I really feel sorry for disturbing your conversation here)

[quote] Ustimenko Igor OneBox CTO wrote: which you came up with here in the topic In fact, the reason is not that you are discussing here, but not even close, but I really feel sorry for disturbing your conversation here) [/quote] I will be sincerely glad if I am wrong in my assumptions. Why the problem arose is not so important. The important thing is that there is no way to promptly signal it to developers. In any case, it is you who has to bear the consequences of troubleshooting the system.... we, integrators, always have the opportunity to "move out" on you, but nevertheless, we also get nuts, but if everything suits you in the current situation, then ok. We don’t change anything, if not, let’s discuss, especially since you have already joined the discussion))

[quote] Ustimenko Igor OneBox CTO wrote: [quote] Farkhshatov Rodion shine-bright.com.ua wrote: [quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: which are loaded without testing on live projects [/quote] problem [/quote] which you came up with here in the topic In fact, the reason is not that you are discussing here, but not even close, but I really feel sorry for disturbing your conversation here) [/quote] 6-8 hours of non-work - a fictitious problem?

[quote] Farkhshatov Rodion shine-bright.com.ua wrote: [quote] Ustimenko Igor OneBox CTO wrote: [quote] Farkhshatov Rodion shine-bright.com.ua wrote: [quote] Pereginyak Alexander Oneboxconsulting (integrator) wrote: which are loaded without testing on live projects [/quote] problem [/quote] which you came up with here in the topic In fact, the reason is not that you are discussing here, but not even close, but I really feel sorry for disturbing your conversation here) [/quote] 6-8 hours of non-work - a fictitious problem? [/quote] a fake problem meant its cause, you yourself came up with a refinement on the basis of which it was born.

issue not resolved I have a question: 1. Can you implement some kind of mechanism that can catch such problems, since now you catch such problems with the help of clients (the problem is that such things break all integrations, and this can lead to bad consequences)? 2. Why didn't cron-minute and cron-hour work, but cron-imap worked (it has some other mechanism)?