Based on New Relic report, about 70% of the time was spent to handle a request consumed by CInlineAction::runWithParams. In fact the Controller::action() took only less than 20% of total time.
At first I thought it might be because of using urlManager to rewrite the request instead of the webserver so I let the rewrite portion to be handled by the webserver and disabled the whole urlManager configuration in my config file.
Based on information provided by Xdebug still there is not much difference.
Just wanted to ask if it's natural for CInlineAction::runWithParams takes that amount of process or there is something wrong with the configuration.
Thanks.
Even though i can't really tell why New Relic reports so much time spent on that method i don't think there's something wrong with your configuration. Yii will create a new CInlineAction object for each call to an inline action (i.e. an action defined inline a controller as an action... function, hence the name). Then runWithParams is then called to run the currently requested action.
So it's quite natural that you see many calls to that method. The only suspicious thing inside that method is that reflection is used to actually call the controller method. But this should not have that dramatic impact on execution time. So i'd probably blame it on the way New Relic does it's measuring.
If you really want to hunt it down, you could add some echo microtime() before each line in that method to find out where the cycles go.
Related
I know that GET is used to retrieve data from the server without modifying anything. Whereas POST is used to add data. I won't get into PUT/PATCH, and assume that POST is always used to update and replace data.
The theory is nice, but in practice I have encountered many situations where my GET calls need to be replaced with POST calls. This is because the response often gets incorrectly cached. Where I work there are proxy servers for security, caching, load balancing, etc., and often times the response for GET calls is directly cached to speed up the call, whereas POST calls never get fully cached.
So for my question, if I have an API call /api/get_orders/month. Theoretically, this should be a GET call, however, the number of orders might update any second. So if I call this API at any moment it may return for example 1000, and calling it just two seconds later should return 1001. However, because of the cache, and although adding a version flag such as ?v=<date_as_int> should ensure that the updated value is returned, there seems to be some caches in the proxy servers that might ignore this.
Basically, I don't feel safe enough using GET unless I know for certain that the data will not be modified or if I know for a fact that the response is always the updated data.
So, would you recommend using POST/GET in the case of retrieving daily/monthly number of orders. And if GET, with all the different and complex layers and server set-ups, how can one be certain that the data is always updated?
If you're doing multiple GET request and something is caching the data in between, you have no idea what it is or how to change it's behavior then POST is a valid workaround.
In any normal situation you would take the time what sits in between your browser and your server, and if there's something that's behaving in a way that doesn't make sense, I would try to investigate and fix that.
So you work at a place where some of that infrastructure exists. Maybe talk to the people that maintain it? But if that's not an option and you just want to find the 'ignore every convention and make my request work'-workaround, then you can use POST.
I'm looking for assistance for the bare minimum code to pull some information from the TheTVDB API (v3).
I've never coded anything to do with APIs before.
I tried to shortcut using TVDBSharper, but that uses asynchronous routines, and tasks, etc. which I just can't get my head around at the moment, given the documentation is for C#, and I clearly don't understand how "await" works in VB.
I've tried searching for API examples, but most are about creating an API.
The first thing TheTVDB API documentation says is:
"Users must POST to the /login route with their API key and credentials in the following format in order to obtain a JWT token."
^ I don't know how to POST. Any examples I've seen are very long and confusing, and mostly in C#.
So (and I apologise for this drivel, but I've tried on and off for months now)…
Could someone please show me the minimal amount of VB.NET code to pull the show name from, for example series ID 73739 (Lost). Hopefully from there, I can start to figure some things out.
I have a valid API Key from the TheTVDB.
Mostly you don't need to understand async/await in any great detail but I was once where you are now, and though I don't claim to be an expert, I did manage to get my head around it like this:
You know how, if you had something that threw an exception and you never caught it:
Sub Main(arguments)
Whatever()
End Sub
Sub Whatever
StuffBefore()
OtherWhateverThrowsException()
StuffAfter()
End Sub
Sub OtherWhateverThrowsException()
StuffBefore()
throw New Exception("Blah")
End Sub
As soon as you threw that exception, your VB thread would stop what it was doing, and wind its way back up through the call stack until it popped out of the main, and crashed to the command line - a matrixy style "return to the source" if you like
Async Await is a bit like that. When there's some method that is going to take a long time to do its work (download strings from tvdb) we could make it sit around doing nothing in our code, having a up of coffee and waiting for TVDB's slow server. This makes things easy to understand because if we sit and wait, we wait 30 seconds, then we get the response, and process the response. Obviously we can't process the response before we get it so we have to sit around and wait for it, and this is always true..
But it'd be better if we could let our thread nip back out the way it came in, "go back to the source", do something else for someone else, and then call it(or another one of its coworkers, we probably don't care) back to carry on working for us when TVDB's server responds. This is what Async Await does for us. Methods that are marked Async are treated differently by the compiler, something like saving your progress on your xbox game. If you reach a point where you want to wait, you can issue the waiting command, the thread that was doing our work performs a savegame, goes off and works for someone else, then when we're ready it comes back, loads the game again and carries on where it left off.
The save game file is manifest as a Task; methods that once upon a time were subs (didn't return anything) should now be Functions that return a Task (a savegame with no associated data). Methods that once upon a time returned something like a string, should now be marked as returning a Task(Of String) - the Task part is to save the state of play (data that VB wants to work with), the string is the data your app wants to work with.
Once you mark something as Async, it needs to contain an Await statement. Await is that SaveYourGameAndGoDoSomethingElseWhileThisFinishes. Typically, while you're awaiting something your program won't have any other stuff it needs the thread to do, so it's not just your Function that calls TVDB's API that needs to Await/be marked Async - every single function in the chain, all the way up and out of your code, needs to be marked as Async, and typically you'll Await at every step of the way back up:
Sub DownloadTVDBButton_Click(arguments)
DoStuff()
End Sub
Sub DoStuff
StuffBefore()
GetFromTVDB()
StuffAfter()
End Sub
Sub GetFromTVDB()
Dim i = 1
GetDataFromTVDBServer() 'wait 30s for TVDB
ParseDataFromTVDB()
End Sub
Sub ParseDataFromTVDB()
End Sub
Becomes:
Async Sub DownloadTVDBButton_Click(arguments) 'windows forms event handlers are always subs. Do not use async subs in your own code
Await DoStuff()
End Sub
Function DoStuffAsync Returns Task
StuffBefore()
Await GetFromTVDBAsync()
StuffAfter()
End Function
Async GetFromTVDBAsync() Returns Task
Dim i = 1
Await GetDataFromTVDBServerAsync() 'go back up, and do something else for 30s
ParseDataFromTVDB()
End Sub
Sub ParseDataFromTVDB() 'downstream; doesn't need to be async/await
End Sub
We switched to using TVB's Async data call, so we await it. When we await, the thread would go back up to the previous function DoStuffAsync. Because we're awaiting that, the thread goes back up a level again into the button click handler. Because we're awaiting that also, the thread goes back up again and out of your code. It goes back to its regular day job of drawing the UI, making it looks like the program is still responding etc. When the TVDB call completes the thread comes back to the point just after it (ready to run ParseData), and it has all the data back from TVDB, and the savegame has been reloaded so everything it knew before/the state is as it was (variable i exists and is 1; you could conceive that it would have been lost otherwise when the thread went off to do something else)
In essence, async/await has allowed us to work exactly as we would have done without it, it's just that it built a little savegame mechanism that meant our thread could go off an do something else while TVDB was busy getting our data, rather than having to sit aorund doing nothing while we waited
It may also help to think of Await as a device that unpacks a save game and gets your data out of it. If a GetSomething() sits for 30s then returns a String you want, then GetSomethingAsync() will instantly return a Task that will (in 30s when the work is done) encloses that same String you want, and Await GetSomethingAsync() will wait until the Task is done then get the string you want out of it
Methods that are named like "...Async" should be thought of as "behave in an asyncronous way". They DON'T have to be marked with the Async modifier; Async is only needed if a method uses the Await word but I'm recommending you use Await on everything that returns a Task (i.e. is awaitable) all the way up and down your call tree. When you get more confident you don't always have to Await SomethingAsync but honestly the overhead of doing so is minimal and the consequences of not doing so are occasionally disastrous. All developers who follow convention always name their stuff ...Async if it behaves in an async way; you should adopt this too, and make sure you name all your Async methods with an"Async" at the end of the name
I don't know how to POST
You don't really need to. The TVDB API has a swagger endpoint; swagger is a way of describing a REST service programmatically so that your visual studio can build a set of classes to use it and provide you with nicely named things. Whipping out a WebClient and manually creating some JSON is very old school/low level
TVDB's swagger descriptor is at https://api.thetvdb.com/swagger.json
You're supposed to be able to right click your project, choose Add... Rest API Client:
,
Paste https://api.thetvdb.com/swagger.json in as the url and pick a namespace (an organizational unit) for all the generated classes to go in.
At the moment something in TVDB's API is causing AutoRest (the tool that VS uses to parse the API spec) to choke but ordinarily it would work out and you'd get a bunch of code (autorest generates c#; you'd be best off generating the c# into a new project and then adding reference to that project from your VB) objects to work with that would do all the POSTing etc for you.
As noted, my VS can't process the TVDB API at the moment and I dont have enough time today to figure out why, but you could sure post a question on AutoRest's github (or on SO) saying "why does https://api.thetvdb.com/swagger.json cause a "Input string not in correct format"" and get some more help
You asked (maybe implicitly) a couple of follow up questions in the comments:
I don't know about REST/swagger (I've heard of it though), and can't see any way to add to the project as you described, and I'm no closer to getting info from TheTVDB. However, it might have have helped me use functions in TVDBSharper. I will just have to try a few things with it. Thanks again
Yes; sorry - I should have been more explicit that "Add REST API client" is only available in a C# project because it relies on a tool that generates C#. This isnt a blocker though - you can just make a C# project and add it to your VB solution alongside your VB project; the two languages are totally interoperable. Your VB can tell your C# what to do
However, there isn't much point in trying at the moment, because the tool that is suppsoed to do it can't handle what TVDB is putting out; my VS can successfully ask the TVDB API to describe itself, but it doesn't seem able to understand the response.
In a nutshell; VS has a bug that means it can't use TVDB API directly, you're best off trying via TvDbSharper. The https://github.com/HristoKolev/TvDbSharper readme has some examples in. They're C# but basically "remove the semicolons and they'll pretty much work in VB"
Now, a bit about the headline terms here, background understanding if you like. API, RESTand swagger are easy enough to explain:
API
An API is effectively a website (in this case run by TVDB), intended for software to consume rather than humans. It takes raw data in and chucks raw data out - unlike a normal website intended for our eyes, nothing about it is presentational in the slightest.
REST
REST as a phrase and a concept is a source of confusion for many and a lot of times you try and read about what REST means and the blogs quickly start getting bogged down with details and make it too complex, with all these funky examples. They kinda forget to explain the REST part because it's come to mean not much at all - it's something so obvious and nondescript that we don't think about it any more.
In essence, something is RESTful if the server doesn't have to remember something about what you did before, in order to service a request you make now - every request stands on its own and can be serviced completely without reference to something else. This is a different workflow to other forms where you might want to change the name of something by issuing a editname('newname') command. What name actually gets edited depends on whether you first did selectshow() or selectactor() and also which show or which actor - a workflow like that means the server has to start remembering whether you selected a show or actor, and what show/actor was selected before it can process the editname() command. If you selected show 123, the edit would edit the name of the show id 123. If you selected an actor 456, the edit name would edit the name of an actor 456
Critically, if you replayed the same editname() at a different time a different thing would get edited because the state of your dialog with the server changes. It's kinda dumb to make the server have to remember all that, for everyone, when really we could push the job of identifying whether we want to name an actor or a show and which show, onto the client
By making it that you have editactorname(123,'Jon wayne') you're transferring all the info the server needs to perform the request; your credentials, the actor id, the new name, the fact that it's an actor name and not a show name. All this goes in the one request, and you can replay this request as many times as you like at any time, and it always has the same effect; things that happened before don't affect it (well.. apart from authentication)
It gets a bit woolly if taken literally - "well if the server doesn't remember anything how does it even remember I changed the name of actor 123, to Jon Wayne so it can service my later request of getactorname(123)?" but that's more about the state of the data in the server, not the state of your interaction with the server. Things that are truly stateless are mostly purely calculatory and not too useful; something somewhere needs to be able to remember something or there is nothing to calculate. Things are rarely completely stateless; even TVDB's API requires you to authenticate first, using a user/password/apikey and then the serverissues a token that becomes your username/password/apikey equivalent for every subsequent request - the server has to start remembering that token, or every time you quote it it will say "can't edit actor name; not authorized". So, yeah.. when viewed holistically something usually has to be rememberd at some point otherwise nothing works. REST things are rarely 100% truly stateless, but mostly they are - and it's really about that "when you want to edit the actor name, send a) that you want to edit actorname, b) what actor, c) what name, d) your credentials to prove youre allowed to" - everything the server needs in the one hit
Swagger
Now called OpenAPI, swagger is a protocol for describing an API: when an api has some actions that take some data, and return some data, it's helpful to know what the actions are called (setactoryearsactive), what type of data they take (date, date), what sort of things you should put in it (the from date, the to date or null if still active), what they return (boolean) and what the return means (true if success, false if not).
If we have a standardized way of describing these things then we can build standard software that reads the standard description of the API and writes a bunch of standard code that uses the API. This is software that writes a description so other software can read it and write software that uses the first set of software. It's an API API.
There is a lot of software here:
The API is software(tvdb),
The thing that generates the description of the API is software (Swagger),
The thing that consumes the description of the API and creates a client is software(AutoRest),
And the thing that uses the client is software (your app).
You could code your app to hit the api directly- the API's just responding
to HTTP requests, which are just text files formatted in a particular way sent to port 80 of the web server that hosts the API. You could write one such request in notepad and use telnet to send it and get a valid response. You could code your app to do it (you were just about to). You could use someone else's library (TvBbSharper) which does it somehow. You could use some software that generates something like TvDbSharper; it reads the description of the api and generates classes for you to use; those classes will make the http requests. Everything can be done at any level; you could write all your apps in assembler, the lowest of the low. It takes ages and it is boring - this is why we use ever higher levels of abstraction.
We make something and then make it do a thousand things and then realize that listing the same code over and over and changing one bit each time is boring, and repetitive and something a computer should do, so we devise ways of making it so software can write the boring repetitive code so that we can do the interesting things.
Swagger and AutoRest are those kind of things; Swagger inspects all the methods, what they take and return and generates a regular consistent description. AutoRest reads it and generates a regular consistent set of client classes. Then the human uses the client classes to do the interesting things. The AutoRest part doesn't work out for us at the moment; it's written by different people than the Swagger team so some differences arise; Awagger describes something and Autorest can't understand it. It will one day I'm sure (in this game of walls and ladders); such is the nature of open source - everyone has a different set of priorities.
Right now we could probably get AutoRest working by finding the one thing it is choking on and removing it. There may be no need; if the TvDbSharper guys have written enough of a set of client classes that you can use TvDbSharper to do all your necessary things. It is thus effectively already the set of client classes AutoRest would have built, maybe more; use TvDbSharper.
The idea behind Swagger and Autorest is that a TvDbSharper shouldn't need to exist: it's a very specific application, only works with tvdb, only works in .net.
If we put effort into making Swagger able to generate a description of any API written in any language, and we put effort into making Autorest able to consume that description and output any language, then we have something more useful than TvDbSharper/no need to TvDbSharper because we can generate something that does the same (of course, specific applications can be superior, just like bespoke tailored suits are superior bt that's another philosophy for another time)
I am interested in loading some system params into the Yii::app()->params array from the database using a CActiveRecord extension called SiteSetting.
Unfortunately I couldn't find much advice online for this, but believe I can place a method in SiteSetting called loadSiteSettingsToAppParams and add the setting...
'onBeginRequest'=>array('SiteSetting', 'loadSiteSettingsToAppParams')
...to the config.
I would like to know if I can only add this onBeginRequest to the Yii::app() somewhere within the SiteSetting class (to keep my code modular) and whether this is a sensible approach.
Thanks in advance.
Just re-read your question now and I'd try to provide answers.
To the question "I would like to know if I can only add this onBeginRequest to the Yii::app() somewhere within the SiteSetting class (to keep my code modular)": the answer is, You're not restricted to just a Class. You could (theoretically) place it anywhere within your application and also in the config.php file.
As to whether it's a sensible approach, it depends on the time it would take to request those settings from the database and whether you're prepared to add that time to your HttpRequest response time. The onBeforeRequest is fired before every HttpRequest and if the loadSiteSettingsToAppParams method consumes lots of time, you're adding that time to your HttpRequest response time.
I'd advise that you fetch those settings once after login and then update them only when they change (the settings are updated). This way, you could place the call to loadSiteSettingsToAppParams in the UserIdentity class and call it after a successful login.
That's just how I'd go about doing this though.
Hope I helped.
The easy & nice way to accomplish this by using a comoponent like SettingComoponent and place in the components directory protected/components then pre load this component in the preload section like this preload => array('log', 'setting', ...). That's it and now you can call this component anywhere you want like Yii:app()->setting->whatever.
I hope this is answer can be useful for you.
I am struggling (in some sense) to determine which HTTP method is more appropriate for rebooting a remote resource: GET or PUT?
On one hand, it seems more semantic to call http://tools.serviceprovider.net/canopies/d34db33fc4f3?reboot=true because one might want to GET a representation of a freshly rebooted canopy.
On the other hand, a reboot is not 'safe' (nor is it necessarily idempotent, but then a canopy or modem is not just a row in a database) so it might seem more semantic to PUT the canopy into a state of rebooting, then have the server return a 202 to indicate that the reboot was initiated and is processing.
I have been reading up on HTTP/1.1, REST, HATEOAS, and other related concepts over the last week, so I am still putting the pieces together. Could a more seasoned developer please weigh in and confirm or dispel my hunch?
A GET doesn't seem appropriate because a GET is expected, like you said, to be "safe". i.e. no action other than retrieval.
A PUT doesn't seem appropriate because a PUT is expected to be idempotent. i.e. multiple identical operations cause same side-effects as as a single operation. Moreover, a PUT is usually used to replace the content at the request URI with the request body.
A POST appears most appropriate here. Because:
A POST need not be safe
A POST need not be idempotent
It also appears meaningful in that you are POSTing a request for a reboot (much like submitting a form, which also happens via POST), which can then be processed, possibly leading to a new URI containing reboot logs/results returned along with a 303 See Other status code.
Interestingly, Tim Bray wrote a blog post on this exact topic (which method to use to tell a resource representing a virtual machine to reboot itself), in which he also argued for POST. At the bottom of that post there are links to follow-ups on that topic, including one from none other than Roy Fielding himself, who concurs.
Rest is definitely not HTTP. But HTTP definitely does not have only four (or eight) methods. Any method is technically valid (even if as an extension method) and any method is RESTful when it is self describing — such as ‘LOCK’, ‘REBOOT’, ‘DELETE’, etc. Something like ‘MUSHROOM’, while valid as an HTTP extension, has no clear meaning or easily anticipated behavior, thus it would not be RESTful.
Fielding has stated that “The REST style doesn’t suggest that limiting the set of methods is a desirable goal. [..] In particular, REST encourages the creation of new methods for obscure operations” and that “it is more efficient in a true REST-based architecture for there to be a hundred different methods with distinct (non-duplicating), universal semantics.”
Sources:
http://xent.com/pipermail/fork/2001-August/003191.html
http://tech.groups.yahoo.com/group/rest-discuss/message/4732
With this all in mind I am going to be 'self descriptive' and use the REBOOT method.
Yes, you could effectively create a new command, REBOOT, using POST. But there is a perfectly idempotent way to do reboots using PUT.
Have a last_reboot field that contains the time at which the server was last rebooted. Make a PUT to that field with the current time cause a reboot if the incoming time is newer than the current time. If an intermediate server resends the PUT, no problem -- it has the same value as the first command, so it's a no-op.
You might want to get the current time from the server you're rebooting, unless you know that everyone is reasonably time-synced.
Or you could just use a times_rebooted count, eliminating the need for a clock. A PUT times_rebooted: 4 request will cause a reboot if times_rebooted is currently 3, but not if it's 4 or 5. If the current value is 2 and you PUT a 4, that's an error.
The only advantage to using time, if you have a clock, is that sometimes you care about when it happened. You could of course have BOTH a times_rebooted and a last_reboot_time, letting times_rebooted be the trigger.
What's the best strategy to use when writing JMeters tests against a web application where the values of certain query-string and post variables are going to change for each run.
Quick, common, example
You go to a Web Page
Enter some information into a form
Click Save
Behind the scenes, a new record is entered in the database
You want to edit the record you just entered, so you go to another web page. Behind the scenes it's passing the page a parameter with the Database ID of the row you just created
When you're running step 5 of the above test, the page parameter/Database ID is going to change each time.
The workflow/strategy I'm currently using is
Record a test using the above actions
Make a note of each place where a query string variable may change from run to run
Use a XPath or Regular Expression Extractor to pull the value out of a response and into a JMeter variable
Replace all appropriate instances of the hard-coded parameter with the above variable.
This works and can be automated to an extent. However, it can get tedious, is error prone, and fragile. Is there a better/commonly accepted way of handling this situation? (Or is this why most people just use JMeter to play back logs? (-;)
Sounds to me like your on the right track. The best that can be achieved by JMeter is to extract page variables with a regular expression or xpath post processor. However your absolutely correct in that this is not a scalable solution and becomes increasingly tricky to maintain or grow.
If you've reached is point then you may want to consider a tool which is more specialised for this sort of problem. Have a look web testing tool such as Watir, it will automatically handle changing post parameters; but you would still need to extract parameters if you need to do a database update but using Watir allows for better code reuse making the problem less painful.
We have had great success in testing similar scenarios with JMeter by storing parameters in JMeter Variables within a JDBC assertion. We then do our http get/post and use a BSF Assertion and javascript do complex validation of the response. Hope it helps