TIBCO Cloud Mashery API Developer Blog

Update:

ATTN Mashery Customers:

Please be advised that the disruption to our standard infrastructure has now been resolved and we are running under normal operations again.  All failed over traffic is once again running on our standard infrastructure (Amazon EC2). 

At approximately 6:33pm PDT, our monitoring services started to report a limited number of issues.  At 6:49pm PDT, the problems became enough to trigger our auto failover services.  Shortly thereafter, Amazon resolved their issues.  Mashery Operations tested the resolution and failed over traffic was routed back to the standard network at 7:36pm PDT. 

Should you wish further information, please contact support@mashery.com or your Client Services Manager.  We apologize for any disruption this may have caused.  

- Mashery Operations

+++

Attn Mashery Customers:

At approximately 6:36pm PDT, we started experiencing sporadic packet loss on our standard infrastructure (Amazon EC2).  All standard and enterprise traffic failed over to our backup alternate infrastructure.  Network DNS entries have 2 minute TTL’s, so some traffic would have experienced timeouts during those minutes.  At this point, traffic is routing well on our alternate infrastructure. 

Premium network customers were not affected.

We will update with more information shortly.

- Mashery Operations

Mashery Clients,

Please be advised that Mashery will have a partial service outage for maintenance purposes this Sunday morning from 12:00am-2:00am Eastern Daylight Time (EDT).   During this period, the following services will be unavailable for routine system maintenance:

1. Client Developer Portals
2. Mashery API

Please note, the Mashery API Proxy Service will continue to operate normally and no developer calls to your API will be impacted. Please contact support@mashery.com or your Client Services Manager with any questions.

Thank You

Please be advised that we will be performing maintenance on our Reporting system early Sunday morning from 3:00am-4:00am Eastern.  The Reporting system will still be available during this time period, however, data updates will be disabled so your reports will be slightly older than normal. 

Thank you for your understanding.  Please send questions to support@mashery.com or contact your Client Services Manager.

Another nice improvement from David on our reporting team was released.  The API Usage summary sparklines have been improved on the Reporting->Overview page to include a count of "Other Responses".  We also modified the hint tips next to each sparkline to better explain what is included in each count. 

Previously, we only displayed "Successful Calls" and "Blocked Calls".  The "Other Responses" sparkline contains all of the other types of call results.  Typically, this is a much smaller number than the successful and blocked calls, but possibly still interesting to you.  It will include call responses such as timeouts, errors, resource not founds, redirects, and so on. 

Here's an example of the new version of this chart:

As always, feedback is welcome!

Our reporting team made a nice little improvement to two of the Reports Dashboard csv extracts today. 

1. Top 5 Error Response Codes: the chart still will show the top 5, but the csv extract now includes data for all error codes and counts in the time period.  Previously, it only extracted the top 5 error responses.
2. Top 5 Method Calls: the chart still will show the top 5 method calls, but the csv extract now includes data for all methods and counts in the time period.  Previously, it only extracted the top 5 method calls.

Let us know what you think.

Mashery Clients,

Please be advised that Mashery will have a partial service outage for maintenance purposes this Saturday, January 10th, from 9:00-9:30pm Eastern Standard Time.   During this period, the following services will be unavailable for routine system maintenance:

1. Client Developer Portals
2. Mashery API

Please note, the Mashery API Proxy Service will continue to operate normally and no developer calls to your API will be impacted. Please contact support@mashery.com or your Client Services Manager with any questions. 

Thank You

We are pleased to announce a major upgrade to the editor in the CMS. While we were upgrading the editor, we were able to sneak in a few other improvements (thanks David!).  The following summarizes some of the changes.  We hope you like them!

1. New Body Editor: we've implemented a new editor in the CMS. You'll see many new action buttons that offer some nice functionality.  For example, there is a set of table management tools so you don't need to write HTML to include tables on your page.  The old HTML view toggle button has been replaced with a pop-up window so you may see both the HTML and resulting page at the same time.  There are a few formatting features also for inserting block indents, subscript/superscript, strikethroughs, etc.  And finally a nice text find and replace feature!

2. More Editing Space: there are two new ways to get more screen real estate for editing your content.  First, the documentation tree can be hidden via toggle.  Second, while editing the page body, you may enter into full screen mode which hides everything but the body editor.

3. File Manager: many people were confused about how to upload files to their site so we pulled the file upload capability out of the page body section and placed it in the action bar.

4. Folder Creation Change: moved the "Create Folder" action out of the "New Item" action list and into the folder viewer page.

Here is a sample screen shot.  Log on and check it out!

We were working with another client recently and found that we had a little weakness around how we provide error message responses back to calling apps for some protocols.  As you may know, our API Proxy Service handles a number of your business rules like rate limits, throttling, security and so on.  When one of these rules fail, or there is generally a technical problem with the call, we respond immediately back to the caller with some type of error response.  And we've always allowed you to customize the error responses to meet your needs. 

As is pretty common in REST APIs, in order to make development a bit easier, many will allow the developer to request a preferred response format.  So if you like looking at XML more than JSON, for many APIs, add a parameter or some indicator of some kind to the call (e.g. "?output=JSON" or "/method.xml") and you get your responses in the format you wish.  Good stuff, but we found a little problem in our app.  The custom error formatting feature only allowed you to serve up one.  So we could respond with JSON, but not XML for example.  So while you were responding with JSON, every once in a while when they got an error, they'd get Mashery XML.  We fixed that today.

Now, you can define what we call an "Error Set".  You can then define the logic for picking the Error Set to respond with based on attributes of the call.  Simple little feature, but one that makes your developer's experience much tighter.  Have a look and contact us at support@mashery.com if you have any questions.

Step 1: Define your Error Sets.  The Error Set Name should be a value that your developers will use in their call.

Step 2: Configure your endpoint. Select the Default Error Set for cases where the developer doesn't ask for something specific.  Then set the Selection Logic for picking the error format that your developer may request.  If your accepting params in the GET or POST of a call, enter the param name.  If you are using a resource type, such as api.mashery.com/search.xml, select the "Resource Type" selection logic and no value is necessary in the Selection Location field. 

Short but sweet posting - reports online.