| Commit message (Collapse) | Author | Age | Lines |
|
|
|
|
|
|
|
|
| |
The HTML/detail should probably be tweaked, but I think we need a way to
show that Gondul is still alive now that we are hiding the time stamp by
default.
Feel free to throw the HTML/CSS away in favor of some other subtle but
noticeable "animation".
|
|\ |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fixes #93
Fixes #92
Fixes #91
The hash logic with FreezeThaw wasn't consistent, which broke a lot of
optimizations. This might be slower (?) but it's consistent which means
we'll make up for it by better cache utilization.
To utilize this I also added delayed JSON parsing (#91) with a big
description of why we want it. This also means that #90 and #89 is a lot
less important. One issue right now is that "time" isn't part of the ETag.
This is both a blessing and a curse. If we are going through time we will
actually use in-app cache (not browser cache - the URL changes). This
makes for nice CPU savings, but ultimately means we can't tell that the
data for two distinct times is identical because we only see one. Oh
well...
On the simpler side, I fixed/improved the clearBox logic to fix the
artifacts. Might need to tune this. I also added some optimizations to
avoid redrawing identical text. This is what pointed out to me that the
hash was broken too: The resizeEvent was triggering all the time, forcing
redraws all the time. This happened because nmsData.switches "updated"
frequently which should only happen if someone adds/moves/removes a switch
or linknet - regardless of time travel. This was caused by the inconsistent
hash causing a false cache miss.
Anyway, if you are still reading this, you almost got a patch with "kjell
magne bondevik uten mellomnavn" as a literal text string... TOO BAD.
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As a result of testing with TG16-data, I was able to find and fix a few
outstanding issues and tune the health map to make more sense.
Also introduced both temperature and dhcp into the health map and info box.
Removed the switchInfo text on the health map unless the score is above a
threshold.
Still more tuning needed. After introducing dhcpInfo it's become even more
important to address #88
|
|
|
|
|
|
| |
- Don't overload the logbook overlay
- Neater date formatting (ugh, I hate that we need to do this)
- Caption for the logbook in the info box
|
|
|
|
| |
Makes it more snug and look the same all over
|
|
|
|
|
|
|
| |
I can never get it all in a single commit.
But I compensate by some times putting multiple unrelated things in the
same commit.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes #69 #11 #5
Introduces nmsTime which unifies the time travel code a bit. It still needs
some work, but this is much better.
All conversion is now done by native JavaScript methods, freeing us from
the hell that is parsing it ourself.
One thing should be added: The backend should discard any now=values that
are not 5-minute intervals. We don't want to kill the cache and the
database by extension.
Still need to re-implement the "replay event" shorthand, but that ties in
to #54
|
|
|
|
|
|
|
| |
- Time in oplog is now localtime and properly padded (00:10, not 0:10)
- Health map instead of combo/aggregated
- Tweak the menu slightly
- Rotate and resize the random switch in guess_placement
|
| |
|
|
|
|
|
|
| |
Fixes #84
It's really stupid simple, but it will work.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
testTree() will make it easier to avoid try-catch'ing because instead of
if (nmsData == undefined || nmsData.snmp == undefined || nmsData.snmp.snmp
== undefined || nmsData.snmp.snmp[sw] == undefined ||
nmsData.snmp.snmp[sw].misc ...)
you can do
if (!testRoot(nmsData,['snmp','snmp',sw,'misc')) { ...
Which is at least slightly less annoying.
|
|
|
|
| |
Knocking stuff off the top of the cpu hogs from a tiny bit of profiling.
|
|
|
|
|
|
|
| |
Specially bad was the foo == bar == undef thing which broke completely.
Also, this makes sure that we don't cap out at 200 if the ping reply is
1000ms...
|
|
|
|
|
|
| |
Fixes broken class list when the description was stuff like "worst:
management info" (where the "info" would end up as a class and change the
look).
|
|
|
|
| |
Management IP up top seems sensible.
|
|
|
|
| |
KEK
|
|
|
|
|
|
|
|
| |
regex search on distro is mainly to include it in extended searches, e.g.:
^(distro0|row21-\d)$
This means "" is now, again, needed to match only one switch.
|
|
|
|
|
|
|
|
|
| |
No need to expose 'switch' id in the API - it's internal to the db.
The entire edit-logic should be cleaned up, but at least this fixes the
immediate problem.
Fixes #74
|
|
|
|
| |
Fixes #76
|
|
|
|
|
|
|
|
|
|
| |
Fixes #82
nms.views is gone, and if the anchor (http://.../#anchor) is a
comma-separated list, "tvmode" will be used.
Also ensures that the anchor already contains the updater tag, it wont be
changed. This means that for tvmode the anchor wont change constantly.
|
|
|
|
|
| |
Mainly for future use. To avoid notifications if we just loaded up
recently, for example.
|
|
|
|
|
|
|
|
|
| |
creates some interesting opportunities, as searches can be combined.
E.g.: '(row\d+-2|distro0|presse) can be used to mark multiple different
switches now. Or '(::1$|^noc$|127.*\.14$) etc. Could probably get rid of
the ""-hack now too. It was mainly meant for the automatic oplog entry on
switch-creation, which can now use "^switch$" instead.
|
| |
|
|
|
|
| |
That was supposed to be a debug thing...
|
|
|
|
| |
Also does ipv6 now.
|
|
|
|
|
|
|
|
|
|
| |
Fixes #80
I should go to bed.
Also: Fixes a tiny little issue where we didn't take into account the
nature if the OPing timeout? Or has it changed? Either way, no longer do we
store thousands of ping replies - per second.
|
|
|
|
|
|
|
|
| |
Fixes #2
Also, since I didn't bother splitting the commits, this tweaks the
information provided from the snmp handler and management handler. Notably
treating IPv4 and IPv6 management addresses as equals.
|
|
|
|
| |
Ooops...
|
| |
|
|
|
|
|
|
|
|
|
|
| |
handlerInfo is getting slightly extensive, but it might be worth it.
This also introduces a "map" handler that .... doesn't handle the map.
If/when namespaces are introduced properly in nms-map-handlers.js, we
should also rename it to nms-data-logic.js or something else that signifies
that this is what makes data useful one way or an other.
|
|
|
|
|
|
| |
Also gets rid of color_from_latency() (?) as it was basically worthless by
now, and adds switch info to the combo-map so you can quickly identify why
a switch is colored a certain way without clicking on it.
|
|
|
|
|
|
|
|
|
| |
Still not very happy with this, but it's inching closer to acceptable every
day.
Should probably redo the whole set_handler() logic with regards to the
anchor to always use the views-logic where a comman-separated list is
acceptable, then get rid of nms.views all together.
|
| |
|
|
|
|
|
| |
You can now use "foo" to find a switch called foo, but not also the ones
called foobar foobar1, etc
|
| |
|
|
|
|
|
|
|
|
| |
Fixes #15
Still lots of things that _should_ be done, but this is the groundwork
needed for a combined map. Theoretically, all map handlers should just
expose getInfo somehow now.
|
|
|
|
|
|
| |
Should get rid of the get_gradient_from_latency entirely.
PS: jocke, you read too many commit logs.
|
|
|
|
|
|
|
|
|
|
| |
Only combines SNMP and ping for now, but the "API" it establishes should do
what we need.
References #15
Before I consider this done we need to add the infomration in the info-box,
and also remove the manual information present there.
|
|
|
|
| |
Needs a lot of work, including the API side, possibly also the schema.
|
|
|
|
| |
Also: onclick handler for log entries.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
It still needs to be integrated with switches/info box, but this is a good
start.
Shows the last 5 log messages in an overlay on the map, with the rest
visible in an oplog tab.
Fixes #20
|
|
|
|
| |
Fixes #41
|
|
|
|
|
|
|
|
|
| |
- Shadow for map title text (color matches background, so only visible when
background interferes with readability)
- SNMP map is private too
- Remove 'map' from a bunch of titles
- Fix z-index and position for map title (z-index is lower than the input
canvas, but higher than everything else).
|
|
|
|
| |
Probably a few more items that needs to be added here.
|