What I see is that there is some change in the email behavior in 6.1. There were some incidents which were setup in 5.6 and It was going as a single mail to all those in the To list.
Now in 6.1 each of them get a different mail with them in the To list. How can I control this? I want to have the same behaviour as we had in 5.6..
Also let me know of any other similar changes btwn 5.6 and 6.1.
Answer by Emerson M. ·
Hi Jibi,
Sorry for the question out of the record.. I see you already have upgraded from 5.6 to 6.1. We are considering to do the same in the short term, Since you have already done this, could you please share your experience with the migration process? after Dynatrace migration have you noticed any performance improvements at all ?.. for example: one of the challenge we experience today with 5.6 is that transaction flow takes long time to show up.. also it's very slow when we try to pull up any purepath..I hope with the new smart analytic engine introduced in version 6.1 all this performance challenge could be overcome.
Appreciate your insight on this
Thanks
Emerson,
I´m not a customer (with experience in that respect). I´m not marketing. I happen to sit next to the engineering guys who implemented exactly the improvements you ask for and, documenting it, I´m pretty deeply involved in it:
1) We split off the task to serve analysis data from the Dynatrace Server and introduced a Frontend Server in 6.0 which gives the Server more headroom and stability and the FS with its own process and memory exclusively caters for the Clients.
2) Check out the new Transaction Flow in the Monitoring Overview you land in in 6.1 which has been optimized for speed and quick initial information.
3) Frontend Server and the caching and fetching algorithm (Smart Analytics Engine in marketing speak) should help you to get analysis data faster, BUT...
Regarding PurePath speed you may want to re-evaluate your approach, not start with PP and possibly a large time frame, but something else and then drill down to the PP data you need. I would like to ask you to consult one of our product experts. Very probably they can show you a better, faster way.
Again, I´m not marketing. It´s just my product knowledge I´m trying to share.
Thanks
G.
Emerson - I do have another issue. This I have not seen any time before in any previous versions. More and more users.. even today I had an incident.. are reporting having difficulties connecting to dt server. It could be proxy, VPN etc. Investigation is still in progress and so far not able to find a single solution that fits all.... Two users who reported the issue were long time users. So ideally it should have worked with the default settings(that they were using earlier with 5.6) for them... SUPDT-6026 is the case I opened.
There seems to be a workaround for this issue. Tested on 2 users who had reported the issue and seems works. Thanks Support team for providing the inputs. Don't test the connection. Just add details and Click OK and Apply. Some bug with test connection with proxy etc it seems.
As this seems to be nailed, let me a bit more explicit on that, not to spread FUD:
There is a problem with Test Connection in Settings menu > Dynatrace Client... > Services vertical tab in the 6.1 Client.
If there is a proxy configured on the OS level, the connection test fails when the Client config is default. (no manual proxy config set in this dialog tab)
SOLUTIONS / workarounds:
Hope this makes it clear
G.
We tested this all on users who were on Proxy and all of them were able to connect successfully as per the solution Guenter mentioned above (In the locally installed Client do not press Test Connection in Settings menu > Dynatrace Server... > Connectivity. Just Apply or OK if settings changed and see if the connection works).
Also for those who started connecting via proxy, I have suggested to go back to the type as 'Default' and not to use the test connection option for now.
Thanks All...
Answer by Jibi U. ·
Guenter - Single email for all in the To list was how we had. People respond to the thread and conversation continues and actions follow. Now, with each person/group getting emails, there can be separate conversations, getting all of them to a single thread etc is difficult. Really NOT a preferred approach.
I was looking at Splunk alerts etc and there also a single mail to all people in the To list. Same as the 5.6 behavior we had.
From looking at the incident emails I get it seems like the behavior hasn't changed, I still have all the recipients in the To field. Could you share a screenshot of an incident configuration and how that email looks for you?
Hi Jibi and Wolfgang,
I just checked with Peter from support that we are seeing the same as Jibi: Mails for an incident get sent separately per recipient!
wolfgang.gottesheim@dynatrace.com, any idea what you could have configured differently that you see the opposite? Because...
Engineering says this has been an intentional change in 6.1 to reduce the number of mails by grouping incidents together (and not recipients).
They also say that alerting groups in the Dynatrace incident rule won´t help, because the addresses will be divided up by incident again.
They are aware of the problem and will think about it in the next version.
BUT we got a SOLUTION for you, Jibi: Please define a group in Exchange / your mail server, and, voilà!
I´ll make a note in the documentation and spread it with support.
Hope that helps
G.
Guenter - I strongly believe we should go back to the earlier behavior. Just as any mail client would send..
Thanks for the temporary solution but as you understand getting these done is not an easy task and the recipients being different based on the incident etc, it makes even more complicated.
Just in case we need to discuss and make sure we are on the same page, I am happy to discuss this over a call and time which works for all. I am based out of Chicago.
I am seeing more mails after this . I generally add myself and the group that I am a part of also in the To.. email list. This has some practical benefits. Only because of this , I could immediate figure out the change.
Hey, Jibi,
thanks for all your input. I do understand.
Yesterday I discussed this with the senior engineer who did the change intentionally and I just deposited your arguments. (He already had thought about it anyway.)
With all the smilies in your reply I think I can try to be funny.
I´ve seen you grow here. Here is what I think people shout out to cheer to somebody:
Jibi, you are The man! U are da man! U-tha-man! (Do you hear the analogy? )
Cheers
G.
Guenter - Thanks. You are the second person to say U-tha-man in the past few months. And actually in my native language it really means.. "the best man".. thanks.
Seems we are both having fun on the job! <s>
Interesting that your name means exactly that. May I ask: what your native language is?
Cheers
G.
Today I got around to put the below in the 6.1 documentation.
I would hope this precisely explains the behavior change and pretty exactly mirrors what you are asking for.
Cheers
G.
You should be aware that, as of 6.1, incident mails are no longer grouped / aggregated by recipients, but by incident.
In other words: Until 6.0 a single mail was sent with all subscribers in the To field when triggered by an incident.
In 6.1, in the attempt to reduce the number of mails, incidents happening in close temporal proximity are grouped together in one mail, sent separately to each subscriber.
So, in 6.1 both goals (less mail and, e.g., reply to group) can be achieved by creating groups on the mail server level.
As this feature may require a considerable amount of work outside the domain, it is under investigation as we try to facilitate things.
Answer by Jibi U. ·
Guenter - I have attached screenshots to the case comparing current and old behavior. Hope it helps.
I should have read the text on the connection screenshot to the end (what I normally do <fg>).
Sorry for asking, but did you check Settings menu > Dynatrace Client > Services vertical tab if there is a proxy configured?
Weird things happen. I brought my iPad yesterday to check web dashboards and nothing worked – only to find out after much head-scratching that I had a 1.5 year old proxy setting in my company WLAN settings. <arrggh>
The iPad forgets the carrier settings with each major upgrade and there was even a bug for this exact scenario in iOS 7, but it remembered the proxy like an elephant! <sigh>
And: What is the problem with one mail with all in the To field vs. mails with one To each? Do I misunderstand?
Thanks
G.
I am still trying to get details from all on the connection issues. No single solution is working for all. Another two people reported issues. Will keep you posted.
Also wondering if there is a way to start the client in debug mode or something with additional logging etc so that it is easier to collect data from these users who are reporting not being able to connect?
Finally I was able to get this working for one user. As far as client settings are concerned, both the web start client and the regular client had the same settings. Web start was working and not the regular. Finally this user had to use connect via proxy by providing the proxy host... extremely slow and connected after some time. This user was always using the default option in 5.6. I never published any other connection types to the company so I am also sure about that being the case for 5.6.
I managed to capture all the details and client logs for this user. Will update SUPDT-6026
Answer by Jibi U. ·
Also more users are reporting some errors connecting to the DT server. so far 3. All previous 5.6 users and same settings... many users are able to connect too..
I already opened a case for this . These were 5.6 users and so ideally it should have worked considering the fact that it was working before. I was wondering if there is a way to increase the timeout values on the frontend server... proxy no proxy, vpn a lot a factors will be at play.. for one user for whom nothing worked, the web start client worked. Yet to do some additional debug on his PC...
SUPDT-6026
Hi Jibi,
re 5.6/ Webstart Clients connecting and stand-alone Clients not: Did you try the IP instead of the host name?
What is the problem with one mail with all in the To field vs. mails with one To each?
Thanks
G.
JANUARY 15, 3:00 PM GMT / 10:00 AM ET