I'm trying to create an alert for what seems a simple alert scenario:
If my logic app is not started for a period of 24 hours I want to
raise an alert for unexpected inactivity.
When I create the alert via the portal I can only select a timewindow of up to 6 hours.
But if I set it up via the Portal API directly then I can configure a timewindow up to 1 day.
My question is why is the portal fixing the window size maximum to 6 hours for an alert? And am I going about this the right way - is there a better way to determine inactivity of certain logic apps? e.g. how could I alert if a logic app didn't trigger for 2 days?
Thanks.
According to the documentation, Alert Rules can be created with a windowsSize between 5 minutes and 1 day in ISO 8601 duration format.
As you mentioned, this is supported via the API, and ARM Templates as well.
why is the portal fixing the window size maximum to 6 hours for an
alert?
I believe this is just a UI limitation, but shouldn't be there.
My other question is am I going about this the right way - is there a
better way to determine inactivity of certain logic apps? e.g. how
could I alert if a logic app didn't trigger for 2 days?
This is an OOTB functionality, I would recommend to use it as long as it meets your requirement. If you needed something else, like a longer period, you would need to implement something custom. You could use the Workflow Runs - List API and check it programmatically.
HTH
Related
Let's say you have a small project. The team has estimated all the tasks as 300 days of effort.
I have 5 developers in the team, and I want MS Project to tell me when the project will complete considering vacations and working schedule of my team member.
In order to do that:
I'm creating a Task "Development" with fixed work "300d", and task type "Fixed Work".
Then I create 5 resources, and specify a 2 week vacation for one of the developers somewhere in the middle of the schedule.
Then I assign my 5 development resources to this task.
The problem is, the 300d distributed evenly to all 5 development resources. And If one of them have a two weeks vacation in between, due to that particular resource the work will be finished 2 weeks later, where other 4 resources are sitting and doing nothing for 2 weeks. Total duration is 70 days.
what I get
What I want to get is: work is distributed accordingly through all 5 resources unevenly in a way that the whole task finishes as earlier as possible taking most of the usable time from all developers.
That's how I would expect it to work. In that particular case I was distributing hours manually.
what i would expect
Is there a possibility in MS Project to do something like this? Or am I doing something wrong?
There are a couple issues with how you are approaching the problem.
1. Rather than just planning out the manpower hours estimated to be needed for the entire project on a single line item, You should plan out the tasks that will need to be done to accomplish "Small Project"
If you discretely plan out the tasks that need to be accomplished to satisfy the scope of "Small project", you can establish dependency (predecessor/successor) relationships between your tasks and figure out what tasks need to be done before you can move on to others. When you do this it will give you a good idea of how long the total duration of the project will take and likely be more accurate than just relying on an estimate based on the manpower hours estimate your developers give you. Find out what tasks they actually need to do, not just how many hours they think the whole project will take them. This will also allow you to plan out the utilization of your resources better because you'll be able to assign specific resources to specific tasks, and not all of your resources need to be on every task.
2. In general I would avoid using the Task Usage form.
I noticed you are altering resources in the task usage form, but unless you are really experienced with Microsoft Project I would avoid ever touching that, as it's really easy to set the period of performance of resources assigned to a task to be different than the actual period of performance of the task itself. This will cause MS Project to behave unusually, and it can be hard for an unexperienced user to understand why. This usually leads to pain and frustration. This leads me to my next bit of advice:
3. If you really want to specify a resource's vacation time, it's better to adjust the calendar associated the resource to exclude those dates as working dates.
In your situation with only 5 resources on your project, this can be fairly easy to do. You can accomplish this 2 different ways (I'll start with the easiest option):
1. You can add resource specific exclusion dates to the default calendar in your project
You can accomplish this by opening the Resource Sheet table and then clicking the Project tab then Change Working Times. If you have the Resource Sheet open instead of the Gantt chart, you can specify the resource that is going to be effected by the exceptions:
In this example you can see that I would be excluding (removing) 8/23/21 thru 9/3/21 as working days for the SW Engineer resource, without needing to change the calendar used by the resource completely.
2. You can completely change the calendar used by particular resources to be different than the default calendar set for the project.
You can accomplish this by going into the Resource Sheet and opening the Base Calendar column:
From here you can assign any calendar that exists in the project to the resource. Of course this means you would need to create the calendars and assign exclusion dates to them.
To create a calendar, click the Project tab then click Change Working Times. Click Create New Calendar on the form that opens up and give it a name:
From there you can add exclusion dates and all that.
Note: In a larger project with many resources, I would recommend not messing with the calendar for the resources at all. It just gets hard to deal with when there are a lot of resources.
The current interval of 1 second and max of 60 seconds is too small and issues may be missed.
When viewing the live metrics stream page of Application Insights the interval on all graphs are 1 second, and it only goes up to 60 seconds. I am trying to use this as a monitoring page to keep an eye on recently released or updated function apps. For this I need to be able to change the interval to view more data at once without having to keep watch on it. Right now if we don't keep watch on it every minute we may miss something important.
I have searched the Microsoft documentation, the git repository, stackoverflow, and various other sites trying to find my answer but the only thing I found was from over 4 years ago and I would hope that this has changed since then.
Live Metrics Stream allows to peek at what's going one right now with 1 second resolution. But it doesn't persist data anywhere. So, data is only stored in UX (browser) and right now only for 60 seconds.
For bigger intervals it might make sense to refer to other Application Insights experiences (including Analytics).
I have a question regarding "Add Calendar By URL" function in Google Calendar:
How often it is updated (most sources I've found says 24h per day). Does caladress.ics?noCache workaround still works?
How it is updated? If I have a large calendar (e.g 2008 - 2016) and add a single event, does Calendar reupload the whole calendar or check for diff? If check for diff, is there any limitations?
Is there any limit to how long events could be? E.g is it possible to set 5 year event?
1. How often it is updated (most sources I've found says 24h per day).Does caladress.ics?noCache workaround still works?
Based from the Google thread, updates may take a few hours for the new information to be parsed and viewable by your users.
Note: It might take up to 12 hours for changes to show in your Google Calendar.
You can use no-cache to indicate that the returned response cannot be used to satisfy a subsequent request to the same URL without first checking with the server if the response has changed. Here is the documentation and example.
2. How it is updated? If I have a large calendar (e.g 2008 - 2016) and add a single event, does Calendar reupload the whole calendar or check for diff? If check for diff,is there any limitations?
Calendar is updated based on how you will implement the "incremental synchronization" of calendar data. It can be Initial full sync or Incremental sync.
Initial full sync is performed once at the very beginning in order to fully synchronize the client’s state with the server’s state. You can optionally restrict the list request using request parameters if you only want to synchronize a specific subset of resources.
While Incremental sync allows you to retrieve all the resources that have been modified since the last sync request. You need to perform a list request with your most recent sync token specified in the syncToken field. Keep in mind that the result will always contain deleted entries, so that the clients get the chance to remove them from storage.
3. Is there any limit to how long events could be? E.g is it possible to set 5 year event?
For the limitation, the Google Calendar API has a courtesy limit of 1,000,000 queries per day. You can see the calendar usage limits here. It is possible to set an event as long as you haven't reached the limit for the number of events you can create.
I have a question about dealing with date & time in my web application. The application will sell monthly subscriptions. It shows dates only when clients buy and cancel subscriptions. Clients can buy additional services in the middle of the month. Application calculates a pro-rata to charge the client until his anniversary date.
I will store dates/times in UTC. It is only for US clients.
I am considering the following options and I would love to get feedback from more experienced developers:
1 - always present dates in EST. I can include a small caption explaining that all subscriptions uses EST. This would be simple as I would not have to deal with clients' timezones. However I am not so sure if clients would be put off by this. Any thoughts?
2 - always present dates in EDT. This would probably not work very well as it would be harder to explain the reason for using it. However I believe it would be simpler to process than EST.
3 - ask for client's timezone information when he signs up for the service and use that information. I don't think this would add too much complication, however I would have to offer them an option to change timezones and I would have to decide what to do with existing subscriptions when there is a change in timezone. If I go with this option I would ask the client to pick the timezone from a drop-down list.
4 - ask for client's location (City and State) and calculate the timezone myself.
5 - try to guess client's timezone based on his IP or another method (ideas???).
Options 3, 4 and 5 would probably be the most user-friendly. Option 1 seems to be the most straightforward to implement.
Sorry for the long post. If you took the time to read it, would you mind taking a little more time and sharing your thoughts and experience?
Thank you.
UPDATE 1 - 9/3/2011 - 17:08 MST
Just found out that PayPal records transactions using PDT and shows them to the client using the client's local timezone they setup when they signed up with PayPal.
I am now inclined to:
1 - show current date using PDT (to align with PayPal) - I will probably change the code to show date & time PDT. Currently I am only showing the date. I believe it will be clearer to the client if I also show the time.
2 - I will not show the anniversary date. I will let PayPal handle that. I will simply state that it is a monthly billing.
3 - When clients add a new service I will calculate the pro-rata using PDT and I will give them a three day grace period to account for timezone differences (thanks to Robert Levy below for suggesting it) and PayPal processing (I don't want to charge them a pro-rata amount if they are only a couple of days from their regular monthly charges).
Any thoughts?
Update 2 - 9/3/2011 - 21:01 MST
Just a quick update. After further research I found out that PayPal does send me a transaction date back. I am not going to show any dates until the client pays at PayPal and I receive confirmation. I will show PayPal's transaction date in the client's receipt.
Sounds like a plan. What do you think?
Just add a grace period of 24 hours from UTC. Easy to code, no extra UI, and unlikely to upset any customers.
my problem is follow:
I have an auctioneer site, in which many different objects will be auctioneerd.
My problem is very simple to clear for an more experience user I thinK. How I can handle business and database logic without opened a site or them?
My problem is to say directly, if nights at 3 no user is on the site, the winner (e.g.) must be set - if a page is opened or not.
So I need some kind of "every 2 seconds, do this method" - without opened a site.
My idea was a sepereate application which uses the same business and database-layer as the asp.net page and let this run at the server. Is that a good or bad idea?
The separate process (scheduled app or Windows service) is the only reliable way you can achieve this.
Using the same BLL and DAL are exactly the right thing to do too.
Check out this article on windows services http://msdn.microsoft.com/en-us/library/aa984074(VS.71).aspx
To let every user query the database every 2 seconds would create unnecessary traffic on your site, which is not a good idea. (users tend to refresh the page many times just before the auction closes anyway)
My thoughts:
Add a date to the auction when it closes.
The last user that places a bid is always the winner, and you can't place bids after 3 am (the date the auction closes), so if you visit the site after 3 am (you can't place a bid and) the winning user is displayed. If somebody opens the site just before 3 am and places a bid after 3 am your business logic should check the date of the bid and deny it.. (also: users might live in different timezones so consider displaying the server time on your site).
Setting the date to 'now' would close the auction immediately.
You can also add javascript to refresh the page if your clock passes the hour or something like that. (or use the number of seconds left before the auction closes in a setTimeout function or metarefresh)