Welcome Guest, you are in: Login

VoIP Innovations Wiki

Image


NAVIGATION
RSS RSS
Main Page
Sales & Marketing
Developer
Porting/Provisioning
Support
Back Office
Titanium Private Label





Search the wiki
»

PoweredBy

Server Failure Notifications

RSS
Modified on 2013/11/25 17:44 by randys Categorized as Support
At VoIP Innovations we are always looking for ways to enhance customer experience and make it easier for you to run your day-to-day operations. One of the ways we believe we can facilitate in easing your day-to-day operations is by providing notifications on critical data. For example, receiving notification that an international spending threshold has been reached can not only save you time and money but it can also alert you to a potential security issue with your server. Another notification we provide is an email to alert you if we are seeing consistent failures on inbound calls to a specific IP address. This article outlines the parameters of when you will receive this email alert and how you should respond to these notifications.

There are many reasons (and corresponding SIP codes) for which inbound calls can fail on your end. While it would be nearly impossible to create a system to alert you to every possible SIP failure we believe it is critical that you are made aware of a failure to respond to inbound INVITEs. If our switch is sending INVITEs for calls to a specific IP address and these INVITEs are being ignored, this points to a serious problem with your server, (or at the very least, an issue that you should address immediately). Our system is designed to send four INVITEs on inbound calls before any kind of failover kicks in, whether that be DID Failover or IP Address Failover. Each INVITE is sent in one second intervals. If you have no failover set up and your switch did not respond to any of the four INVITEs we send, our switch will end the call with a Release Cause:INVITE Response Timeout (SIP:504). While you are certainly encouraged to have failover in place, failover can take several seconds to be effective. At minimum, you have added four seconds of PDD for each IP address that fails to respond to INVITEs as our system must send four INVITEs before advancing to the next IP address in a group. To mitigate these types of issues for you, we have designed a notification system to send you an email alert if your server fails to respond to an INVITE to at least 50% of inbound call attempts.

This notification system runs constantly but you will not be bombarded with emails every few seconds. Instead, you will be sent an email once the 50% failure threshold has been reached. To reiterate: You will only receive notifications in regards to INVITE Response Timeouts. You will NOT receive failure notifications for any other SIP rejections. Once you have been sent an email notification regarding the failure, you will not receive another email for a period of one hour. If you have not corrected the problem or the IP address continues to be non-responsive to INVITEs (again with a 50% failure rate), another email will be sent after an hour has passed. The emails are sent to the Technical Contact on the account. If no there is no Technical Contact configured, the email will be sent to the Main Contact on the account. (You can configure a Technical Contact from the Account Tab (Account->Update Contact Information) in the Back Office.) Be aware that the notifications are on a per endpoint basis so you will receive multiple notifications if multiple IP addresses fail to respond. While you will not receive an email alert for an hour once it's been sent the first time, you may receive additional emails for other IPs if they are also not responding.

If you are receiving the server failure notifications you have several options to correct this. Our first recommendation is to start with the server behind the IP address causing the problem to find out why it isn't responding to inbound calls. If the IP address is no longer active, submit a removal request through the Back Office. If you are unable to immediately correct the problem with the server, you can either change the priority of IP addresses (set to an IP address that WILL accept the call) or set the IP address to inactive. Only Endpoint Groups that have multiple IP addresses in them can set priority and only Endpoint Groups with multiple IP addresses in them can be set to inactive. For example, if you have one IP address in a group there will be no edit button to make the IP inactive (there must be at least one active IP address in any group that has DIDs assigned to it). One thing to keep in mind in regards to IP priority is that the priority is set in order of appearance. You may see numbers listed under the priority but you can ignore them. We will send calls to IP addresses in the order that they appear from the top down. Obviously if you have three IP address in a group and the second one is set to 'Inactive', that one would be ignored in the order calls are sent.

While these emails are labeled as "Server Failure" notifications, it is possible that the issue does not lie with the server or switch itself. VoIP Innovations bases this notice off the fact that INVITEs are not being responded to, meaning the INVITEs are leaving our network and being sent to the IP address you have configured in the Back Office. It is certainly possible that the packets have dropped in transit. The way you can eliminate your server as being the issue is to check your switch logs. If your switch logs have no record of the calls we notified you about, it is possible the INVITEs are not reaching you. This indicates a network problem somewhere in the path. Our best recommendation is to contact our support team when the issue is occurring (after you get the email and verified your switch logs) and we can perform traceroutes to your network and provide you the data that you can provide your ISP to look into further. Please note that it is important to contact our support team as quickly as possible as things on the internet change quickly. If you wait a day or two after receiving the notifications before contacting support, there will be little information we can provide if the issue has ceased as traceroutes at that point will likely be clear.

We have not allowed for opting out of these email alerts. The main reason is that if you are receiving the emails, there is a problem that needs addressed on your side. Even if you no longer need an IP address you should still remove it from your account. INVITE Response Timeouts use up system resources on our side so having you address the situation is beneficial to both you and VoIP Innovations. If you refuse to address the problem and do not want the email alerts we can only suggest you create an email rule to filter them although we prefer you correct the problem causing the alerts.

The actual email you will receive will contain the total number of calls, the total number of calls that received no response, the Endpoint Group number and Group name, and the IP address the calls were sent to. The DIDs called and the numbers that called them are also included in the email. The notification system will be checking calls to your endpoints on a 15 minute interval. As an example, if we send 100 calls to an IP address in a 15 minute period and 51 receive no response, you will receive the email alert. If this pattern continues for the next 15 minutes, you will not receive an email alert until an hour has passed. If we send 100 calls in a fifteen minute period and 49 receive no response, you will not receive the notification as this falls below the 50% failure rate. Additionally, this will be on a per endpoint basis so if you have multiple servers not responding, you will receive multiple alerts (each containing information pertaining to that specific endpoint).

Here is a sample of the email alert that we will send:
Dear Valued Customer,

We are sending you this email as a courtesy because we thought you would like to know we have seen a failure rate of at least 50% on inbound traffic to one of your IP addresses. Our system monitors your traffic and if at least a 50% failure rate is detected, we send this email. Below are the details:

Endpoint Group: 566
Group Name: Test Group
IP Address: 51.63.122.167
Total Calls: 27
Failed Calls: 27

DID(s) Called: 7244192266 called from 14128242201

Judging by this data, it appears that the server may be experiencing sporadic network issues. We suggest investigating what may be causing this issue. If this server has been taken down permanently or is no longer in use, please log in to the Back Office and remove it from your Endpoint Groups.

You will not receive any further notifications for this IP address for at least 1 hour. Please contact our support or see our wiki article for more information on these notifications: http://wiki.voipinnovations.com/Server-Failure-Notifications.ashx

Below are a few screenshots pertaining to changing priority of IP addresses and setting an endpoint as 'Inactive'.

To change priority or active status, hover the 'Endpoints' Tab and click 'My Endpoints'

Image

Click the 'Expand' button to open edit options

Image

Grayed out checkboxes show current status of IP addresses (active IPs will have a check in the box)

Image

Change priority by clicking the 'Up/Down' arrows or click the Pencil icon to open an additional Edit Window to set to Inactive

Image

Use the check box and click the 'Update' button to set Active/Inactive status

Image

ScrewTurn Wiki version 3.0.5.600. Some of the icons created by FamFamFam.