Since your doc is not leading you into standard treatments for Cluster let us throw some ideas at you.
Imitrex pills are the least effective form of this med for Cluster. They are too slow acting. The injection form works within 10-15 minutes and will give, for most people, total relief.
You mention a seeking for prevention, the triggers. This is not a clear issue with Cluster with the exception of alcohol which must be avoided when you are in an active headache period. Sometimes solvents will trigger but the list is not as uniform as you might expect.
The best approach to prevention is outlined in the following protocol which is widely used in the U.S. Suggest you print it out and share with the doc.
-----
Headache. 2004 Nov;44(10):1013-8.
Individualizing treatment with verapamil for cluster headache patients.
Blau JN, Engel HO.
Background.-Verapamil is currently the best available prophylactic drug for patients experiencing cluster headaches (CHs). Published papers usually state 240 to 480 mg taken in three divided doses give good results, ranging from 50% to 80%; others mention higher doses-720, even 1200 mg per day. In clinical practice we found we needed to adapt dosage to individual's time of attacks, in particular giving higher doses before going to bed to suppress severe nocturnal episodes. A few only required 120 mg daily. We therefore evolved a scheme for steady and progressive drug increase until satisfactory control had been achieved. Objective.-To find the minimum dose of verapamil required to prevent episodic and chronic cluster headaches by supervising each individual and adjusting the dosage accordingly. Methods.-Consecutive patients with episodic or chronic CH (satisfying International Headache Society (IHS) criteria) were started on verapamil 40 mg in the morning, 80 mg early afternoon, and 80 mg before going to bed. Patients kept a diary of all attacks, recording times of onset, duration, and severity. They were advised, verbally and in writing, to add 40 mg verapamil on alternate days, depending on their attack timing: with nocturnal episodes the first increase was the evening dose and next the afternoon one; when attacks occurred on or soon after waking, we advised setting an alarm clock 2 hours before the usual waking time and then taking the medication. Patients were followed-up at weekly intervals until attacks were controlled. They were also reviewed when a cluster period had ended, and advised to continue on the same dose for a further 2 weeks before starting systematic reduction. Chronic cluster patients were reviewed as often as necessary. Results.-Seventy consecutive patients, 52 with episodic CH during cluster periods and 18 with chronic CH, were all treated with verapamil as above. Complete relief from headaches was obtained in 49 (94%) of 52 with episodic, and 10 (55%) of 18 with chronic CH; the majority needed 200 to 480 mg, but 9 in the episodic, and 3 in the chronic group, needed 520 to 960 mg for control. Ten, 2 in the episodic and 8 in the chronic group, with incomplete relief, required additional therapy-lithium, sumatriptan, or sodium valproate. One patient withdrew because verapamil made her too tired, another developed Stevens-Johnson syndrome, and the drug was withdrawn. Conclusions.-Providing the dosage for each individual is adequate, preventing CH with verapamil is highly effective, taken three (occasionally with higher doses, four) times a day. In the majority (94%) with episodic CH steady dose increase under supervision, totally suppressed attacks. However in the chronic variety only 55% were completely relieved, 69% men, but only 20% women. In both groups, for those with partial attack suppression, additional prophylactic drugs or acute treatment was necessary. (Headache 2004;44:1013-1018).
=======================================
SLOW-RELEASE VERAPAMIL
Dr. Sheftell applauded the protocol for verapamil used by Dr. Goadsby and colleagues, which entailed use of short-acting verapamil in increments of 80 mg. “This method was suggested by Lee Kudrow, MD, 20 years ago as an alternative to slow-release verapamil,” Dr. Sheftell noted.
“I would agree with using short-acting verapamil, rather than the sustained-release formulation, in cluster headache,” he said. “I prefer the short-acting formulation with regard to ability to titrate more accurately and safely. My clinical experience anecdotally demonstrates improved responses when patients are switched from sustained-release verapamil to short-acting verapamil.”
Dr. Goadsby agreed that his clinical experience was similar. “There are no well-controlled, placebo-controlled, dose-ranging studies to direct treatment. This is one of those areas where clinicians who treat cluster headache have to combine what modicum of evidence is available with their own clinical experience,” Dr. Sheftell commented.
====
See the PDF file, below, for a current overview of treatments for Cluster.
====
A rapidly actling abortive, an alternative to Imitrex injection:
Headache 2001 Sep;41(8):813-6
Olanzapine as an Abortive Agent for Cluster Headache.
Rozen TD.
Department of Neurology, Jefferson Headache Center/Thomas Jefferson University Hospital, Philadelphia, Pa.
OBJECTIVE: To evaluate olanzapine as a cluster headache abortive agent in an open-label trial. BACKGROUND: Cluster headache is the most painful headache syndrome known. There are very few recognized abortive therapies for cluster headache and fewer for patients who have contraindications to vasoconstrictive drugs. METHODS: Olanzapine was given as an abortive agent to five patients with cluster headache in an open-label trial. THE INITIAL OLANZAPINE DOSE WAS 5 MG, AND THE DOSE WAS INCREASED TO 10 MG IF THERE WAS NO PAIN RELIEF. THE DOSAGE WAS DECREASED TO 2.5 MG IF THE 5-MG DOSE WAS EFFECTIVE BUT CAUSED ADVERSE EFFECTS. To be included in the study, each patient had to treat at least two attacks with either an effective dose or the highest tolerated dose. RESULTS: Five patients completed the investigation (four men, one woman; four with chronic cluster, one with episodic cluster). Olanzapine reduced cluster pain by at least 80% in four of five patients, and TWO PATIENTS BECAME HEADACHE-FREE AFTER TAKING THE DRUG. Olanzapine typically alleviated pain within 20 minutes after oral dosing and treatment response was consistent across multiple treated attacks. The only adverse event was sleepiness. CONCLUSIONS: Olanzapine appears to be a good abortive agent for cluster headache. IT ALLEVIATES PAIN QUICKLY AND HAS A CONSISTENT RESPONSE ACROSS MULTIPLE TREATED ATTACKS. IT APPEARS TO WORK IN BOTH EPISODIC AND CHRONIC CLUSTER HEADACHE.
PMID 11576207 PubMed
--------------------------------------------------------------------------------
Olanzapine has a brand name of "Zyprexa" and is a antipsychotic. Don't be put off by this primary usage. Several of the drugs used to treat CH are cross over applications, that is, drugs approved by the FDA for one purpose which are found to be effective with unrelated conditions--BJ.
====
A very good, general introduction to Cluster:
Cluster headache.
From: Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

or

(Orphanet Journal of Rare Diseases)
[Easy to read; one of the better overview articles I've seen. Suggest printing the full length article--link, line above--if you are serious about keeping a good medical library on the subject.]
Leroux E, Ducros A.
ABSTRACT: Cluster headache (CH) is a primary headache disease characterized by recurrent short-lasting attacks (15 to 180 minutes) of excruciating unilateral periorbital pain accompanied by ipsilateral autonomic signs (lacrimation, nasal congestion, ptosis, miosis, lid edema, redness of the eye). It affects young adults, predominantly males. Prevalence is estimated at 0.5-1.0/1,000. CH has a circannual and circadian periodicity, attacks being clustered (hence the name) in bouts that can occur during specific months of the year. ALCOHOL IS THE ONLY DIETARY TRIGGER OF CH, STRONG ODORS (MAINLY SOLVENTS AND CIGARETTE SMOKE) AND NAPPING MAY ALSO TRIGGER CH ATTACKS. During bouts, attacks may happen at precise hours, especially during the night. During the attacks, patients tend to be restless. CH may be episodic or chronic, depending on the presence of remission periods. CH IS ASSOCIATED WITH TRIGEMINOVASCULAR ACTIVATION AND NEUROENDOCRINE AND VEGETATIVE DISTURBANCES, HOWEVER, THE PRECISE CAUSATIVE MECHANISMS REMAIN UNKNOWN. Involvement of the hypothalamus (a structure regulating endocrine function and sleep-wake rhythms) has been confirmed, explaining, at least in part, the cyclic aspects of CH. The disease is familial in about 10% of cases. Genetic factors play a role in CH susceptibility, and a causative role has been suggested for the hypocretin receptor gene. Diagnosis is clinical. Differential diagnoses include other primary headache diseases such as migraine, paroxysmal hemicrania and SUNCT syndrome. At present, there is no curative treatment. There are efficient treatments to shorten the painful attacks (acute treatments) and to reduce the number of daily attacks (prophylactic treatments). Acute treatment is based on subcutaneous administration of sumatriptan and high-flow oxygen. Verapamil, lithium, methysergide, prednisone, greater occipital nerve blocks and topiramate may be used for prophylaxis. In refractory cases, deep-brain stimulation of the hypothalamus and greater occipital nerve stimulators have been tried in experimental settings.THE DISEASE COURSE OVER A LIFETIME IS UNPREDICTABLE. Some patients have only one period of attacks, while in others the disease evolves from episodic to chronic form.
PMID: 18651939 [PubMed]
===
The problems with chronic use of pain meds are known
known and are not worth risking in the presence of several other effective, safe abortives.