Help link
is a context-sensitive link to this Help topic.
Help for Foxhound 1.1.3832a
Table of Contents [RisingRoad]
The Foxhound Options page lets you change a variety of Foxhound settings and perform a variety of Foxhound-specific tasks.
Settings in this page are global in nature: any changes you make will apply to all Foxhound users and sessions running on this copy of the Foxhound engine.
5.4 Connection Sampling Threshold
The New Menu link opens the Foxhound Menu page in a new browser window or tab.
The
Help link
is a context-sensitive link to this Help topic.
The About link opens the About Foxhound page in a new browser window or tab.
The button hides this Help frame from view, and brings it back.
The checkbox Show the Help frame on the right side by default changes Foxhound's behavior whenever a new page is opened: if checked, Foxhound will initially show the right-hand Help frame. If unchecked, Foxhound will not initially show the right-hand Help frame. It doesn't affect pages that are already open, just what happens for new pages.
You must click on the button if you want the checkbox setting value to be saved (not this fake button, but the real one on the Foxhound Options page :)
The Send Alert emails: checkbox specifies whether or not you want emails to be sent whenever one or more alert criteria are met. You can change this value for individual target databases by using the Alerts Criteria page.
Use Email address(es) for Alerts: field to specify one or more email addresses, separated by semicolons, that are to receive the email alerts. You can change this value for individual target databases by using the Alerts Criteria page.
Click on the Use HTML in emails: checkbox if you want the email alerts to be formatted using HTML. You can change this value for individual target databases by using the Alerts Criteria page.
The Host[:port] for URLs in emails: field may be used to specify where the Foxhound web server is located relative to the computer receiving the emails, so that HTML links pointing to Foxhound History pages from inside the email alerts can find the right web pages. The default is localhost which works if the emails are being received on the same computer that's running Foxhound. If you start Foxhound on some port other than 80, you can specify a value like localhost:12345. If Foxhound is running on a different computer you'll have to change localhost to that computer's domain name or IP address; e.g., xyz.com:12345. You can change this value for individual target databases by using the Alerts Criteria page.
Click on Use SMTP to send emails radio button to specify that SMTP will be used to send all alert emails, instead of MAPI.
The SMTP Sender: holds the email address of the sender; e.g., abc@xyz.com will appear in the email "from"
field as Foxhound Alert
The SMTP Server: is a server domain name or IP address of the SMTP server; e.g., smtp.xyz.com
The SMTP Port: specifies the TCP/IP port to use for SMTP messages.
The SMTP Timeout: specifies how long to wait, in seconds, before giving up
on an attempt to send an email. The Foxhound Database Monitor sents email alerts
asynchronously so a long timeout doesn't adversely affect the monitoring process.
However, the Send Test Email button does work synchronously; i.e., it waits until the email is send,
so you can use it to determine how long the timeout should be.
The SMTP Authorization User Name: holds the user name when the
SMTP server requires authentication.
The SMTP Authorization Password: holds the password when the
SMTP server requires authentication.
Strong encryption is used to
store this value in the Foxhound database, and the value is not redisplayed on this page.
In other words, this field is input-only; just because it's empty doesn't necessarilymean
there's no password stored.
Click on Use MAPI to send emails radio button to specify that MAPI will be
used to send all alert emails, instead of SMTP.
The MAPI User Name: holds the MAPI login name; e.g., postmaster@xyz.com
The MAPI Password: holds the MAPI login password.
Strong encryption is used to
store this value in the Foxhound database, and the value is not redisplayed on this page.
In other words, this field is input-only; just because it's empty doesn't necessarilymean
there's no password stored.
You must click on the button if you
want the email settings to be saved.
The button may be used to erase
the values currently stored in SMTP Authorization User Name and the SMTP Authorization Password.
This button is necessary if you switch from using an SMTP server that requires authorization to one
that doesn't; Foxhound doesn't redisplay the values entered in those fields so you can't just
"blank them out" to erase them, you have to press this button.
The button may be used to check that
the email settings are correct.
The Test result: field will display OK if the test email was sent ok, or
an error message if there was a problem. However, you should also check that the
email was received at the other end; just because a message was sent doesn't mean
it will actually arrive.
The Connection sampling threshold... radio buttons let you specify when (and if)
Foxhound will stop recording detail information about individual connections. Here
are your choices:
You must click on the button if you
want the new Connection sampling threshold... value to be saved.
The new value takes effect immediately, for all sampling sessions, as soon as you save it.
Here are the main reasons you might want to set a threshold other than All:
The threshold is an "all or nothing" setting. I.e., if you set a threshold other
than zero or All, Foxhound will stop gathering any detail information about any
connections when the ConnCount database property exceeds the threshold. When
ConnCount drops back to the threshold or lower, Foxhound will start gathering
detail information about all the connections again.
Note that ConnCount is retrieved before the detail connection information so there may be a
change in the number of connections by the time the detail information is retrieved. For example, if
you set the threshold to 100, you may see detail information about more than 100 connections for a
short period if new connections are being created.
The Connection timeout: radio buttons
control how long the Foxhound Monitor will wait for a response
when trying to connect to a target database:
1 minute, or never timeout.
You must click on the button if you
want the new Connection timeout: value to be saved.
Most connection attempts either succeed or fail right away,
and the Foxhound Monitor normally keeps trying to connect if it fails so
that sampling can begin or resume as soon as a connection is
possible.
Sometimes, however, an attempt to connect to a target database
located far away on a WAN will take a very long time to
receive a response if the connection fails. In some of these
cases, the Foxhound engine is blocked from doing other work
while waiting to receive the failed response, and this causes
problems if the Foxhound Monitor keeps repeating this long wait
while trying to monitor other databases at the same time.
The connection timeout facility is provided to stop the Foxhound
Monitor from continuously trying to connect to a database when the
failure-to-connect response takes a long time. When a timeout
occurs, Foxhound will stop trying to connect just as if you
clicked on Cancel Request or Stop Sampling, and it will not
try to connect again until you tell it Start Sampling or the
Timeout Retry
period has elapsed.
If you get a lot of false timeouts, try the 1 minute setting.
If you are only monitoring one database, the never timeout
is probably best: Foxhound's problems with long waits for
failure-to-connect responses are not
important, and sampling will begin as soon as the Foxhound
Monitor is able to connect.
See also 5.4 Timeout Retry.
The Timeout retry: radio buttons control how long the Foxhound Monitor
will wait after sampling is stopped by a connection timeout
before trying to connect to the target database and start sampling again:
10 minutes, or 1 hour, or never retry.
You must click on the button if you
want the new Timeout retry: value to be saved.
When a timeout occurs, sampling is stopped and Foxhound no longer
attempts to connect to the target database.
Normally, when sampling is stopped, it can only be restarted manually.
A connection timeout, however, is a special case: the timeout retry interval
is a mechanism to automatically restart sampling without manual intervention.
For more information about connection timeouts, see
5.3 Connection Timeout.
Foxhound's internal purge process frees up space
in the Foxhound database by deleting old data. The purge process
is automatically launched every 20 minutes, 24 hours a day, starting at 00:00,
and can also be launched manually via the Start Purge button. The purge process
runs independently from all other Foxhound processes, including the Foxhound Monitor
processes and the web service that displays this Foxhound Options page.
Note that the Foxhound database will not shrink in size when data is deleted; the freed space
is simply made available for future needs. In fact, the Foxhound database file may temporarily
grow in size during the purge process to accommodate growth in the checkpoint log as
rows are deleted. This checkpoint log space will be released when the Foxhound database is stopped and
restarted, but the space freed up by the purge process will not be released. This behavior is
a characteristic of all SQL Anywhere databases.
When the Enable automatic purge schedule: checkbox is checked, the purge process
that is automatically launched every 20 minutes will proceed to delete old data. When
this checkbox is unchecked, the purge process will still be launched every 20 minutes
but it won't do anything; in effect, the scheduled purge is disabled.
The Purge all sample data: radio buttons specify
how soon old server, database and connection-level Monitor sample data will be deleted:
after 1 week, after 1 month, after 1 year or never purge anything.
The "after 1 day / 1 week / ..." calculations for Purge all sample data:
are based on the most recent successful sample, not the current timestamp.
That means the most recent "1 day / 1 week / ..." worth of successful samples will be preserved
if sampling is turned off or stops for some other reason. It also means that if there are
no successful samples, nothing is purged.
Note that "all sample data" means just that: if a sample satisfies the "after 1 day / 1 week / ..."
calculation it will be deleted. It doesn't matter it the sample is an Alert, or is associated with a Peak or Alert,
or if it is a "lost" or "failed" sample (target database not found, sampling stopped, Foxhound stopped,
etc.) Also, when a sample associated with a Peak is deleted, the associated entry on the
Peak line will lose its hyperlink attribute.
The Purge uninteresting connection data: radio buttons specify
how soon uninteresting connection-level Monitor sample data will be deleted:
after 1 week, after 1 month, after 1 year or never purge anything.
To have any effect the Purge uninteresting connection data: setting should be more
aggressive (shorter) than the Purge all sample data: setting.
If you specify a longer setting for Purge uninteresting connection data: than
for Purge all sample data: you will see the following message:
In other words, "all sample data" encompasses "uninteresting connection data".
A connection is defined as "uninteresting" when all the following conditions are true for
the previous and current samples recorded by the Foxhound Monitor for that connection:
the connection was not blocked and did not hold any locks then or now (BlockedOn and LockCount = 0), no new request has been started for this connection (LastReqTime and LastStatement were the same then and now), the server wasn't busy servicing this connection (ReqStatus is empty or 'Idle' then and now,
and the ApproximateCPUTime, Commit and Rlbk properties have not changed then or now), and the connection hasn't done any real work (RollbackLogPages = 0 then and now).
When...
the Purge uninteresting connection data: setting is satisfied, and the previous connection-level sample is not associated with a Peak or Alert, then the previous connection-level sample is deleted.
The "after 1 day / 1 week / ..." calculations for Purge uninteresting connection data:
are based on the most recent successful sample, not the current timestamp.
That means the most recent "1 day / 1 week / ..." worth of connection-level samples will be preserved
if sampling is turned off or stops for some other reason.
Note that "lost" or "failed" samples (target database not found, sampling stopped, Foxhound stopped, etc.) are not affected by the Purge uninteresting connection data: setting because those samples don't have any connection-level data to be deleted.
The Purge speed: radio buttons specify how aggressively the purge process will pursue its goal of deleting old data. Choosing a lower number may improve the performance of other processes including the Foxhound Monitor itself. Choosing a higher number may ensure that the purge process will keep up with the workload and prevent the Foxhound database from growing in size at all or too quickly. It's up to you to determine the right setting, and that depends primarily on the performance of the computer running Foxhound and the number of connections to the target databases being monitored; i.e., a faster computer will support a higher Purge speed: setting, and a large number of connections to target databases will require a higher Purge speed: setting to keep the Foxhound database under control.
You must click on the button if you
want any of the above settings to be saved. New settings won't affect a purge process that's
already running, just the next one to start.
The button starts the purge process immediately. Only one purge process at a time will run; a duplicate process will simply stop by itself.
The button will stop a purge process that is currently running. It may take a moment for your request to take effect; the purge process will continue running until it reaches a commit point and then stop.
The button forces the Foxhound engine to perform a checkpoint so that all of the purge-related statistics (row counts, etcetera) shown on this page are up to date. This checkpoint is executed on the Foxhound database, not any of the target databases.
The button forces the data on this page to be recalculated and redisplayed, in particular the Purge Run and Foxhound DBSPACE reports.
The Purge Run report shows the last 10 runs of the purge process, in reverse order with the most recent run at the top.
The Foxhound DBSPACE report shows information about the files comprising the Foxhound database.
If Foxhound encounters a run-time exception, it may store
diagnostic information in an internal table. Foxhound also
stores some informational messages in this table.
The button
lets you display the most recent diagnostic data in reverse
order by timestamp.
The Export diagnostics to file: field and the
button let
you write all the
diagnostic data, sorted by timestamp. If the file exists,
it will be overwritten.
The button
lets you delete the diagnostic data from the Foxhound database.
If all three Diagnostics buttons are disabled, it means there is no diagnostic data to display.
Note: If you upgrade Foxhound to a new build, the data upgrade process will copy all
the existing diagnostic data from the old database to the new one.
The
button will stop the Foxhound database engine.
The button will
refresh the Foxhound Options page with new data.
The button will
close the Foxhound Options browser window or tab.
[Top]
[Top]
5.4 Connection Sampling Threshold [Top]
5.5 Connection Timeout
[Top]
In cases like this, it may be better to run the Foxhound
engine on a computer that is located physically close to the target
database server and communicates over a LAN rather than WAN. For
more information on this topic see the Help section
Introduction and Setup - Running Foxhound as a Service.
5.6 Timeout Retry
[Top]
5.7 Purge Process
[Top]
Warning: The "Purge all sample data:" interval will take precedence over the "Purge uninteresting connection data:" interval.
5.8 Diagnostics
[Top]
5.9 Stop - Refresh - Close
[Top]
The Stop Foxhound Engine button is disabled when the Foxhound database is running on a
computer that is separate from the one running the browser.