2. Close Rates (closed calls only)
| Time | Incidents | % | Cumulative Total |
| 4 h | 1 | 1.41 | 1.41 |
| 12 h | 1 | 1.41 | 2.82 |
| 24 h | 0 | 0 | 2.82 |
| 48 h | 3 | 4.23 | 7.05 |
| 72 h | 1 | 1.41 | 8.46 |
| 1 Week | 2 | 2.82 | 11.28 |
| 2 Weeks | 3 | 4.23 | 15.51 |
| 4 Weeks | 40 | 56.34 | 71.85 |
| >4 Weeks | 20 | 28.17 | 100.02 |
| Total Incidents | 71 |
That doesn't count the 5 TARs I still have open; one from May. 84% of my TARs are closed in four weeks or more. Let me say that again; FOUR WEEKS OR MORE. Each TAR cost me approximately $2500.
So, Oracle Community, did I get adequate return on my investment?
(Oh, and BTW, I love how my cumulative total adds up to 100.02%)
4 comments:
Jeff,
there are at least 2 problems in this kind of statistic:
1) the better the DBA is, the worse are the numbers. (under the asumption, a good DBA can solve the 4-hr-problems alone)
2) these numbers doesn't say anything where the time was spent during the TAR: waiting on customer? waiting on support? waiting on development? Unfortunately I do not know an 10046 event for TARs. (maybe someone should raise an enhancement request?)
I cannot say if oracle support is it's money worth, there is just no option.
I'm just sure, Oracles strategy is the best for Oracle, not for the customers.
Jeff
That's not dissimilar to the sort of stats that I had at my last place of employment. (so I like Martin's first suggestion). I imagine however that you are excluding from your calculation
1) any patches or upgrades applied during the year including the CPU program
2) any issues that you resolved using metalink.
If you include both of the above it's probably < $2000 a pop :(
Niall
I've just gone very near the 100K magic number for overall support.
Ours comes in two chunks, courtesy of an Oracle accounting incapable of coping with an upgrade halfway through a maintenance contract...
Anyways, looking at your numbers and looking at the paltry amount of my calls shames me completely: I am definitely not getting back any value!
:)
One of the things I've long ago given up on is letting metalick handle a call for more than a week. I usually only open severity 1 calls: if it can't be solved within a week, then the workaround gets in and we forget about it.
The way things work at Oracle support, it's hopeless to try and squeeze an emergency patch out of "product development" or whatever those bozos are called nowadays.
Along the years the number of calls I've had cancelled due to "won't be fixed" or "no patch available until version xx.x.x.x.x, kindly upgrade" is staggering...
But the regular patch releases and the list of fixes for each patch release have paid handsome dividends, I'm now more or less as patched up as 10.2.0.3 can get.
Still got to do the partitioning option patch shuffle, later in 2009.
In all: value?
Hmmm, define value first?
What gets left out from this type of analysis is how many times one does not even bother opening up a service request and/or trying to force a recognition out of oracle for a bug because one has realized the ( probable ) substantial wait time.
One has only to be forced around the repeated re-assigns of a sev 2 and/or sev 1 to various people in time zones around the block to ( unfortunately ) become somewhat immunized how less than optimally effective contacting oracle to help resolve a problem might be.
Agreeing with Martin there is not a viable alternative.
Post a Comment